Dissertationen zum Thema „Behaviour discovery“

The first chapter investigates the price and volatility impacts produced by block trades in an inter-market environment with different microstructures. A sample of European cross-traded securities is employed to investigate whether large trades executed on the foreign market (London Stock Exchange's SEAQ-I market) produce any impacts on the securities' home markets and analyse whether different market microstructures matter. The price impact in the home markets is detected before the large trade is executed on SEAQ-I and proceeds in a protracted fashion, implying that substantial pre- and post-positioning is undertaken by London market makers through the home markets. The new equilibrium price on the home market is reached before the trade information is published on SEAQ-I. Large trades are also found to cause higher price volatility in auction trading systems than in a hybrid market microstructure. The second and third chapters analyse the formation of quoted and effective spreads and their components in three different market microstructures. The results show that quoted and effective spreads generated by a hybrid system (Deutsche Borse's IBIS system) are lower than those generated by both the pure auction system (Paris Bourse's CAC system) and the dealership system (London Stock Exchange SEAQ market). Traders on a hybrid mechanism face the lowest costs and this result holds even when we control for (a) the level of market concentration in liquidity provision, and (b) company-specific news. However, the adverse selection component of the spread is significantly higher in an auction trading system compared to both the dealership and the hybrid trading system. This fifth chapter investigates (a) whether, in a hybrid trading mechanism, voluntary market makers provide a higher level of price stabilisation than limit order traders even if they do not have any obligation to keep orderly markets, (b) the strategic interactions between the limit order book and market makers, and (c) the behaviour of the order flow at times of price uncertainty. We analyse these issues using high frequency data from the London Stock Exchange which has adopted a hybrid market microstructure. We find that prices on the dealership system track the security's true value more efficiently. The dealership system can transact higher volumes with lower price volatility. This evidence suggests that market makers provide price stabilisation, even if they have no binding obligation to do so, thus improving the market's quality. In terms of trading behaviour, we find that in a hybrid trading mechanism, traders are not encouraged to provide liquidity on the order book through limit orders as price uncertainty increases. Instead orders migrate to the dealership system for execution.

Polyoxopalladates (POPds) are an emerging class of unconventional polyoxometalates (POMs) which are particularly interesting due to the catalytic properties of palladium. Until recently, POPds were limited to low nuclearity {Pd12} and {Pd15} archetypes. However, in 2012, the giant {Pd84}Ac macrocycle was discovered, becoming the largest of this family to date. One of the major barriers to the discovery of POMs in general is that the chemistry relies heavily on crystallisation. The presence of POMs in solution is often overlooked if no crystalline products emerge. Furthermore, their behaviour in solution is not well-understood. This thesis expands on the family of POPd macrocycles and explores the supramolecular behaviour of the POPd macrocycles using solution analysis techniques such as SEC-HPLC, ESI-MS and NMR spectroscopy. In the first section of work, a new synthetic route to making {Pd84}Ac is reported, whereby palladium acetate is substituted for sodium acetate, allowing the carboxylate source to be added separately. This new synthetic route was investigated to explore the formation of novel macrocycles by varying the acetate ligand. Initial results showed this to be possible by substituting acetate for propionate. This yields a smaller {Pd72}Prop macrocycle, which was characterised crystallographically. An SEC-HPLC approach was then adopted to screen many small-scale reactions for the formation of other large palladium macrocycles, using a wide variety of carboxylate sources. This led to the discovery of {Pd84}Gly, an exact structural analogue of {Pd84}Ac. IMS-MS was employed to assess the relative sizes of the new macrocycles in solution, confirming the presence of {Pd72}Prop and {Pd84}Gly. The structure of {Pd84}Gly was later confirmed crystallographically. Having shown the screening method to be successful in detecting novel POPds in solution, this approach was expanded on. In the second section, SEC-HPLC was used to screen many small-scale reactions containing different combinations of polyphosphate solutions and carboxylate ligands, in the presence of Pd(NO3)2. SEC-HPLC indicated the presence of a variety of species larger in size than {Pd84}Ac, however further analysis using ESI-MS was not able to identify these products, most likely due to their decomposition under the mass spectrometry conditions, or poor separation in the SEC column. Over this course of this work, two smaller POPd clusters containing the squarate ligand were isolated: {Pd11} and {Pd8}. In the third section, two mass spectrometry techniques were employed to analyse the real-time self-assembly of the {Pd84}Ac, {Pd72}Prop and {Pd84}Gly macrocycles. Analysis was carried out using two methods: SEC-HPLC-MS and direct injection of a desalted sample (Desalted-MS). Initial results showed the loss of acetate ligands when using SEC-HPLC-MS, which was confirmed by comparing identical crystalline samples of {Pd84}Ac analysed with each technique. The {Pd72}Prop macrocycle was found to appear in solution after five days, and exhibits building units analogous to those seen in the assembly of {Pd84}Ac, suggesting a similar build-up mechanism to {Pd84}Ac. Likewise, {Pd84}Gly was found to appear in solution after four days and also exhibits the same analogous building units. Furthermore, a {Pd30} species was observed in SEC-HPLC-MS studies of the {Pd84}Gly reaction, indicating a possible template or smaller macrocycle which has not yet been isolated. In the final section, the molecular encapsulation potential of {Pd84}Ac was explored using 1D and 2D NMR techniques. Using 1H NMR spectroscopy, interactions were observed between {Pd84}Ac and a series of polyamine guests, indicating a fast exchange regime between the free and encapsulated guest molecules. Subsequent 1H DOSY NMR spectroscopy showed the diffusion coefficient of 1,8-diamionooctane matched that of {Pd84}Ac when the compounds were mixed, confirming the strong host-guest interaction. Following this, the [Ru(bpy)3]Cl2 (Ru-bpy) complex was selected as a potential guest within the {Pd84}Ac cavity due to its suitable size and photochemical properties. 1H NMR spectroscopy showed a significant shifts in the Ru-bpy 1H signals in the presence of {Pd84}Ac, indicating a fast exchange interaction. 1H DOSY NMR spectroscopy revealed the coefficient of the Ru-bpy complex was found to decrease dramatically when in the presence of {Pd84}Ac.

The eukaryotic interphase nucleus is well organised and the genome positioned non-randomly. Nuclear structure is an important regulator of genome behaviour and function. Genome organisation and nuclear structure are compromised in diseases such as cancer and laminopathies. This study was to find out and to determine if there is any functional relationship between nuclear structure and genome mis-organisation in cancer cells. I have assessed the presence and distribution of specific nuclear structural proteins (A-type, B-type lamins and its receptor LBR, many of their binding proteins such as MAN1, LAP2α, LAP2, and Emerin and other nuclear proteins (PML, Nucleolin, and Ki67) using indirect immunofluorescence. From this study, it is found that the nuclear structure of breast cancer cells is often altered. The most severely affected proteins are the nuclear lamins B1 and B2 and they found as large foci within the nucleoplasm with little LBR expression to localise the lamin B. I also assessed the chromosome positioning (HSA 7, 10, 11, 14 and 17) and gene positioning (AKT1, CCND1, HSP90AA1, EGFR, ERRBB2/HER2 and PTEN) in breast cancer cell lines (T-47D, GI-101, Sk-Br-3 and BT-474) and in normal breast cell lines (MCF-10A) using 2D-FISH technique. I also assessed the position of the genes in nuclei and correlated with gene expression using qRT-PCR. Breast cell lines have treated with a drug named lovastatin and it was found that the cells have restored LBR expression and localisation of lamin B, leading to altered gene positioning and changed expression of breast cancer genes. Since the drug (lovastatin, 12 μM/48 hours) affects the prenylation as a post-translation modification process and lamins B biosythensis, it is found that B-type lamins and its receptor expression and distribution were improved and increased in expression by 2-fold in expression levels in the most affected cells (T-47D, and BT-474) compared to the normal cells (MCF-10A) and these cells also showed abnormal nuclei and dead cells. When analysing the nuclear positioning of the genes (AKT1, HSP90AA1 and ERRBB2/HER2), it is found that AKT1 was positioned periphery in BT-474 and T-47D cells and interiorly in the normal cells (MCF-10A) before treatment whereas the same gene was positioned periphery in T-47D and MCF-10A cells and interiorly in BT-474 after treatment with lovastatin. It is also found that HSP90AA1 was positioned periphery in MCF-10A and T-47D cells and interiorly in BT-474 cells before and after treatment (no change). Moreover, ERRBB2/HER2 gene was positioned periphery in T-47D and BT-474 cells and interiorly in MCF-10A cells before treatment whereas the same gene was positioned periphery in MCF-10A and T-47D cells and interiorly in BT-474 after treatment with the same drug. Regarding LMNB1, LMNB2, and LBR genes, the study focussed only on their expression levels and no work has done on their chromosome positioning as well as gene position before and after treatment. These three genes were over expressed when assessed by measuring the relative and fold changes in expression. Therefore, it is suggestive that 2D-FISH experiment to assess their localisation and their specific chromosome territories is required. The results shown in this thesis demonstrate the importance and roles of nuclear architecture specifically nuclear lamins and the integral nuclear membrane proteins (B-type lamins and LBR) in mediating correct genome organisation and function. The breast normal (immortalised cells) and cancerous cell lines showed different nuclear structures as lamin B affect the position of specific target chromosomes and genes. These results will strength the finding that the nuclear lamina is a significant nuclear structure which associates, organises, and regulates numerous vital nuclear processes and the stability of the genome.

A series of abuse scandals in recent years have stimulated debate regarding appropriate ways to care for people with intellectual disabilities and challenging behaviour, with policy emphasising the need to move beyond reactive and aversive approaches. This thesis presents a qualitative study of the use of Discovery Awareness; a non-aversive video-analysis based approach to challenging behaviour which aims to improve relational aspects of care. Fieldwork was undertaken over a 10 month period in two Assessment and Treatment Units on a single site in the UK. During this period, 7 Discovery Awareness sessions were filmed and 12 interviews were conducted with staff who had experience of attending DA sessions. The filmed DA sessions were analysed using Conversation Analysis (CA) whilst the interview data explored staff accounts and perceptions of using DA in practice. The findings demonstrate that method integrity is an interactionally achieved phenomenon accomplished by the chairperson establishing interactional frameworks, shaping the conditional relevance of turns in line with the aims of the session, and through signalling transgressions in situ. There was no one interactional rule that applied to the whole DA session, making participant transgressions probable. Stance and interpretation were central actions related to video analysis in the session. Participants in DA utilised vague language to reduce the ‘degree of liability’ that a speaker may face. Participants also employed experiential formulations when interpreting the patients’ behaviour. This both minimises the basis for disagreement and enables participants to bring to the surface tacit dimensions of knowledge regarding the patient. These skills may be valuable in care environments, where staff have to make decisions based on signs interpreted from patients in the moment. The conversation analytic work also uncovered the prevalent use of ‘imagined constructed inner dialogue’, in which staff members talk as the patients’ inner voice. This formulation was used to perform a variety of actions such as evidencing participants’ stances, gist formulating the turn of another speaker, and to imagine alternative courses of events from the patient’s point of view. The use of imagined constructed inner dialogue also enabled speakers to give voice to non-aligning turns which offer alternative assessments without the need for prefacing non-alignment and minimising disagreement commonly found in every day conversations. Whilst imagined constituted dialogue fulfils an epistemic function, it also enabled participants to fulfil one of the remits of a DA session; to try to see the world from the patient’s point of view. By speaking as the patient, staff take an empathetic stance and affiliate with the patient’s imagined experience whilst keeping the patient’s voice central, and debatable, in the sessions. The interview findings suggest that DA is perceived as a valuable tool in helping staff reflect on the behaviour and personhood of the patient, as well as their own role in interactions with the patient. Whilst a variety of positive outcomes were reported, there was no consensus about how or whether these factors resulted in a reduction in challenging behaviour. However DA was perceived as having a beneficial effect on the culture of the units and in promoting recognition of personal and collective skills relating to understanding and exploring the patients’ behaviour. Possible impediments to implementation were also identified through analysis of the sessions, staff interviews, and through a literature review focusing on the factors which promote or impede the use of methods in ID practice. The findings contribute to existing conversation analytic work on studying intellectual disability services from an interactional perspective. They also build on work examining how methods are implemented interactionally. This has relevance for understanding how ID methods are implemented in real world situations, and for ID policy and practice.

Comme le titre pourrait le suggérer, l’objectif principal de cette thèse est de fournir une meilleure compréhension de la tâche de la découverte de sous-groupes à travers la théorie de l’ordre. La découverte de sous-groupes (Subgroup Discovery - SD) est la tâche automatique dont le but est la découverte d’hypothèses intéressantes dans les bases de données. Autrement dit, étant donnée une base de donnée, l’espace de recherche de toutes les hypothèses que l’analyste voudra tester ainsi qu’un moyen formel pour évaluer la qualité de ces hypothèses ; la tâche automatique de la découverte de sous-groupe s’efforce de trouver les meilleurs hypothèses quant à ces trois paramètres. Afin d’élaborer des algorithmes efficaces et efficients pour cette tâche, il est important de comprendre les propriétés des espaces de recherche d’une part et les propriétés de la mesure de qualité d’autre part. Dans cette thèse, nous étendons l’état de l’art par: (i) fournir une vue unifiée sur les espaces d’hypothèses derrière la tâche de découverte de sous-groupes en utilisant la théorie de l’ordre, (ii) proposer l’espace d’hypothèses de conjonctions d’inégalités linéaires dans les bases de données numériques ainsi que différents algorithmes permettant de les énumérer et (iii) proposer un algorithme anytime - fournit progressivement des résultats - pour la tâche particulière de fouille de sous-groupe discriminants dans les bases de données numériques. Ce dernier fournit des garanties sur la qualité des sous-groupes extraits même si l’algorithme est interrompu
As the title of this dissertation may suggest, the aim of this thesis is to provide an order-theoretic point of view on the task of subgroup discovery. Subgroup discovery is the automatic task of discovering interesting hypotheses in databases. That is, given a database, the hypothesis space the analyst wants to explore and a formal way of how the analyst gauges the quality of the hypotheses (e.g. a quality measure); the automated task of subgroup discovery aims to extract the interesting hypothesis w.r.t. these parameters. In order to elaborate fast and efficient algorithms for subgroup discovery, one should understand the underlying properties of the hypothesis space on the one hand and the properties of its quality measure on the other. In this thesis, we extend the state-of-the-art by: (i) providing a unified view of the hypotheses space behind subgroup discovery using the well-founded mathematical tool of order theory, (ii) proposing the new hypothesis space of conjunction of linear inequalities in numerical databases and the algorithms enumerating its elements and (iii) proposing an anytime algorithm for discriminative subgroup discovery on numerical datasets providing guarantees upon interruption

Agricultural commodity derivatives were first introduced in South Africa in 1996 after the deregulation of the former marketing system. In the context of its proposed functions, namely price discovery and risk management, the question arose as to whether the futures market developed over time to performed its role efficiently. According to the Efficient Markets Hypothesis (EMH) an efficient market is one that accurately incorporates all information available at any point in time. The purpose of the research was to address the issue of price discovery efficiency, firstly, focusing on the weak-form methodology. Secondly, considering the behaviour of futures prices over time, the study addressed the concern of anomalies in daily returns – phenomena contradictory to the EMH by implication. Thirdly, as a means of defining the sources of inefficiency, the role of scheduled public information and its impact on futures prices was examined. Therefore, the primary objective of the research was to investigate and identify the main components of agricultural futures market inefficiency within the unique price formation structure of South African grain markets. The assessment of this problem is important in terms of evaluating the growth and development of the futures market for different grain commodities to date. The Exchange needs to review rules and regulations on a frequent basis in order to ensure proper functioning at all times especially in the case of a relatively new and fast growing market. The study contributed to the knowledge of understanding the price adjustment process and its implications for market efficiency in the context of the three grain markets considered. The weak-form efficiency was tested using a co-integration based model. Analysing daily spot and futures prices of white maize, yellow maize, and wheat, results indicated that all three markets were efficient and unbiased. Non-parametric tests revealed the significant presence of day-of-the-week and turn-of-the-month effects in the futures returns of the three commodities. Further non-parametric analyses suggested a high degree of uncertainty in futures returns around scheduled agricultural and macroeconomic information release dates also contributing significantly to the identified anomalies. It was concluded that (1) the markets’ ability to anticipate the contents of future information to be released, (2) the current skewed size distribution of broking members, (3) the significant role of the R/$ exchange rate in the price formation process of South African grains and, therefore, (4) the relationship to and influence of the broader economy enhanced the return effects (anomalies) creating opportunity for profitable arbitrage. This conclusion was mainly attributed to South Africa’s status as a price-taker in the world grain complex as well as the relatively short existence of the local agricultural futures markets.

This thesis proposes a collection of analytical and computational methods for inferring an underlying social structure of a given population, observed only via timestamped occurrences of its members across a range of locations. It shows that such data streams have a modular and temporally-focused structure, neither fully ordered nor completely random, with individuals appearing in "gathering events". By exploiting such structure, the thesis proposes an appropriate mapping of those spatio-temporal data streams to a social network, based on the co-occurrences of agents across gathering events, while capturing the uncertainty over social ties via the use of probability distributions. Given the extracted graphs mentioned above, an approach is proposed for studying their community organisation. The method considers communities as explanatory variables for the observed interactions, producing overlapping partitions and node membership scores to groups. The aforementioned models are motivated by a large ongoing experiment at Wytham woods, Oxford, where a population of Parus major wild birds is tagged with RFID devices and a grid of feeding locations generates thousands of spatio-temporal records each year. The methods proposed are applied on such data set to demonstrate how they can be used to explore wild bird sociality, reveal its internal organisation across a variety of different scales and provide insights into important biological processes relating to mating pair formation.

This paper seeks to find whether the spectrum auctions in Sweden have been conducted efficiently and if there is a de facto model that suits all auctions. The efficiency is conditions that emphasise truthful bidding, price discovery and limits collusive behaviour. The paper compares three different auction models used in Sweden, a beauty contest used in the allocation of 3G spectrums, and the auction model selected for the upcoming 5G spectrum auction. The auction models are as follows: first and second-price sealed-bid auction, SMRA and CCA. We found that beauty contests should not be used in any spectrum allocation as it did not meet the criteria of efficiency outlined in this paper. The first-price sealed-bid auction is not a suitable format for spectrum auctions. According to the theory, it generates equivalent revenues on average as the second-price format, which shows a higher degree of efficient allocation. We found that depending on the blocks sold, both SMRA and CCA can result in somewhat efficient results, but they are not suitable for a single object auction. We found that no de facto auction format is suitable for every spectrum auction to be conducted in the future, but instead that the auction format is dependent on the characteristics of the individual auctions.

Code clone Discovery Based on Functional Behavior by Ronald M Krawitz 2012 Legacy programs are used for many years and experience many cycles of use-maintenance-use-maintenance-use-etc. Source code or source code functionality is frequently replicated within these programs when it is written, as well as when it is maintained. Over time many different developers with greater or lesser understanding of the source code maintain the source code. Maintenance developers, when they have limited time or lack understanding of the program, frequently resort to short cuts that include cutting and pasting existing code and re-implementing functionality instead of refactoring. This means a specific functionality is often repeated several times, sometimes using different source code. Blocks of replicated source code or source code functionality are called code clones. Removing code clones improves extensibility, maintainability, and reusability of a program in addition to making the program more easily understood. It is generally accepted that four types of code clones exist. Type-1 and Type-2 code clones are comparatively straightforward to locate and tools exist to locate them. However, Type-3 and Type-4 code clones are very difficult to locate with only a few specialized tools capable of locating them with a lower level of precision. This dissertation presents a new methodology that discovered code clones by studying the functional behavior of blocks of code. Code Clone Discovery based on Functional Behavior (FCD) located code clone by comparing how the blocks of code reacted to various inputs. FCD stimulated the code blocks with the same input patterns and compared the resulting outputs. When a significant portion of the outputs matched, those blocks were declared to be a code clone candidate. Manual analysis confirmed that those blocks of code were code clones. Since FCD discovered code clones based on their black-box behavior, the actual source code syntax was irrelevant and manual inspection further confirmed FCD located code clones that included Type-3 and Type-4 code clones which are frequently excluded from code clone detection tools. FCD recognized the code clones regardless of whether or not they use identical code, similar code, or totally dissimilar code. This new technique allows for an improvement in software quality and has the potential to significantly reduce the cost of software over its lifetime.

L'acquisition et la modélisation de connaissances ont été abordés jusqu'à présent selon deux approches principales : les êtres humains (experts) à l'aide des méthodologies de l'Ingénierie des Connaissances et le Knowledge Management, et les données à l'aide des techniques relevant de la découverte de connaissances à partir du contenu de bases de données (fouille de données). Cette thèse porte sur la conception d'un processus d'apprentissage conjoint par l'être humain et la machine combinant une approche de modélisation des connaissances de type Ingénierie des Connaissances (TOM4D, Timed Observation Modelling for Diagnosis) et une approche d'apprentissage automatique fondée sur un processus de découverte de connaissances à partir de données datées (TOM4L, Timed Observation Mining for Learning). Ces deux approches étant fondées sur la Théorie des Observations Datées, les modèles produits sont représentés dans le même formalisme ce qui permet leur comparaison et leur combinaison. Le mémoire propose également une méthode d'abstraction, inspiée des travaux de Newell sur le "Knowledge Level'' et fondée sur le paradigme d'observation datée, qui a pour but de traiter le problème de la différence de niveau d'abstraction inhérent entre le discours d'un expert et les données mesurées sur un système par un processus d'abstractions successives. Les travaux présentés dans ce mémoire ayant été menés en collaboration avec le CSTB de Sophia Antipolis (Centre Scientifique et Technique du Bâtiment), ils sont appliqués à la modélisation de l'activité humaine dans le cadre de l'aide aux personnes âgées maintenues à domicile
Knowledge acquisition has been traditionally approached from a primarily people-driven perspective, through Knowledge Engineering and Management, or from a primarily data-driven approach, through Knowledge Discovery in Databases, rather than from an integral standpoint. This thesis proposes then a human-machine learning approach that combines a Knowledge Engineering modelling approach called TOM4D (Timed Observation Modelling For Diagnosis) with a process of Knowledge Discovery in Databases based on an automatic data mining technique called TOM4L (Timed Observation Mining For Learning). The combination and comparison between models obtained through TOM4D and those ones obtained through TOM4L is possible, owing to that TOM4D and TOM4L are based on the Theory of Timed Observations and share the same representation formalism. Consequently, a learning process nourished with experts' knowledge and knowledge discovered in data is defined in the present work. In addition, this dissertation puts forward a theoretical framework of abstraction levels, in line with the mentioned theory and inspired by the Newell's Knowledge Level work, in order to reduce the broad gap of semantic content that exists between data, relative to an observed process, in a database and what can be inferred in a higher level; that is, in the experts' discursive level. Thus, the human-machine learning approach along with the notion of abstraction levels are then applied to the modelling of human behaviour in smart environments. In particular, the modelling of elderly people's behaviour at home in the GerHome Project of the CSTB (Centre Scientifique et Technique du Bâtiment) of Sophia Antipolis, France

Avec la prolifération rapide des plateformes de données qui récoltent des données relatives à plusieurs domaines tels que les données de gouvernements, d’éducation, d’environnement ou les données de notations de produits, plus de données sont disponibles en ligne. Ceci représente une opportunité sans égal pour étudier le comportement des individus et les interactions entre eux. Sur le plan politique, le fait de pouvoir interroger des ensembles de données de votes peut fournir des informations intéressantes pour les journalistes et les analystes politiques. En particulier, ce type de données peut être exploité pour l’investigation des sujet exceptionnellement conflictuels ou consensuels. Considérons des données décrivant les sessions de votes dans le parlement Européen (PE). Un tel ensemble de données enregistre les votes de chaque député (MPE) dans l’hémicycle en plus des informations relatives aux parlementaires (e.g., genre, parti national, parti européen) et des sessions (e.g., sujet, date). Ces données offrent la possibilité d’étudier les accords et désaccords de sous-groupes cohérents, en particulier pour mettre en évidence des comportements inattendus. Par exemple, il est attendu que sur la majorité des sessions, les députés votent selon la ligne politique de leurs partis politiques respectifs. Cependant, lorsque les sujets sont plutôt d’intérêt d’un pays particulier dans l’Europe, des coalitions peuvent se former ou se dissoudre. À titre d’exemple, quand une procédure législative concernant la pêche est proposée devant les MPE dans l’hémicycle, les MPE des nations insulaires du Royaume-Uni peuvent voter en accord sans être influencés par la différence entre les lignes politiques de leurs alliances respectives, cela peut suggérer un accord exceptionnel comparé à la polarisation observée habituellement. Dans cette thèse, nous nous intéressons à ce type de motifs décrivant des (dés)accords exceptionnels, pas uniquement sur les données de votes mais également sur des données similaires appelées données comportementales. Nous élaborons deux méthodes complémentaires appelées Debunk et Deviant. La première permet la découverte de (dés)accords exceptionnels entre groupes tandis que la seconde permet de mettre en évidence les comportements exceptionnels qui peuvent au sein d’un même groupe. Idéalement, ces deux méthodes ont pour objective de donner un aperçu complet et concis des comportements exceptionnels dans les données comportementales. Dans l’esprit d’évaluer la capacité des deux méthodes à réaliser cet objectif, nous évaluons les performances quantitatives et qualitatives sur plusieurs jeux de données réelles. De plus, nous motivons l’utilisation des méthodes proposées dans le contexte du journalisme computationnel
With the rapid proliferation of data platforms collecting and curating data related to various domains such as governments data, education data, environment data or product ratings, more and more data are available online. This offers an unparalleled opportunity to study the behavior of individuals and the interactions between them. In the political sphere, being able to query datasets of voting records provides interesting insights for data journalists and political analysts. In particular, such data can be leveraged for the investigation of exceptionally consensual/controversial topics. Consider data describing the voting behavior in the European Parliament (EP). Such a dataset records the votes of each member (MEP) in voting sessions held in the parliament, as well as information on the parliamentarians (e.g., gender, national party, European party alliance) and the sessions (e.g., topic, date). This dataset offers opportunities to study the agreement or disagreement of coherent subgroups, especially to highlight unexpected behavior. It is to be expected that on the majority of voting sessions, MEPs will vote along the lines of their European party alliance. However, when matters are of interest to a specific nation within Europe, alignments may change and agreements can be formed or dissolved. For instance, when a legislative procedure on fishing rights is put before the MEPs, the island nation of the UK can be expected to agree on a specific course of action regardless of their party alliance, fostering an exceptional agreement where strong polarization exists otherwise. In this thesis, we aim to discover such exceptional (dis)agreement patterns not only in voting data but also in more generic data, called behavioral data, which involves individuals performing observable actions on entities. We devise two novel methods which offer complementary angles of exceptional (dis)agreement in behavioral data: within and between groups. These two approaches called Debunk and Deviant, ideally, enables the implementation of a sufficiently comprehensive tool to highlight, summarize and analyze exceptional comportments in behavioral data. We thoroughly investigate the qualitative and quantitative performances of the devised methods. Furthermore, we motivate their usage in the context of computational journalism

Recent advances in sensor technologies and the growing interest in context- aware applications, such as targeted advertising and location-based services, have led to a demand for understanding human behavior patterns from sensor data. People engage in routine behaviors. Automatic routine discovery goes beyond low-level activity recognition such as sitting or standing and analyzes human behaviors at a higher level (e.g., commuting to work). The goal of the research presented in this thesis is to automatically discover high-level semantic human routines from low-level sensor streams. One recent line of research is to mine human routines from sensor data using parametric topic models. The main shortcoming of parametric models is that they assume a fixed, pre-specified parameter regardless of the data. Choosing an appropriate parameter usually requires an inefficient trial-and-error model selection process. Furthermore, it is even more difficult to find optimal parameter values in advance for personalized applications. The research presented in this thesis offers a novel nonparametric framework for human routine discovery that can infer high-level routines without knowing the number of latent low-level activities beforehand. More specifically, the frame-work automatically finds the size of the low-level feature vocabulary from sensor feature vectors at the vocabulary extraction phase. At the routine discovery phase, the framework further automatically selects the appropriate number of latent low-level activities and discovers latent routines. Moreover, we propose a new generative graphical model to incorporate multimodal sensor streams for the human activity discovery task. The hypothesis and approaches presented in this thesis are evaluated on public datasets in two routine domains: two daily-activity datasets and a transportation mode dataset. Experimental results show that our nonparametric framework can automatically learn the appropriate model parameters from multimodal sensor data without any form of manual model selection procedure and can outperform traditional parametric approaches for human routine discovery tasks.

Substance use disorders are serious health concerns in the United States. Furthermore, the National Survey on Drug Use and Health reports a continuous increase in substance use disorders in the United States during the last 10 years. However, there are not many effective pharmacotherapeutics available for substance use disorders. The current dissertation is focused on research aimed at discovering pharmacotherapeutics for substance use disorders. First part of dissertation focused on discovering methamphetamine (METH) use disorder therapeutics targeting specific mechanism of METH action on dopaminergic neurons. The second part of dissertation focused on opioids and cocaine use disorder therapeutics targeting rewarding pathway commonly activated by opioids and cocaine. With respect to METH, it induces release of dopamine (DA) in neuronal terminals by interacting with the vesicular monoamine transporter-2 (VMAT2) and DA transporter (DAT). VMAT2 inhibitors have been found by our research group to decrease METH-evoked DA release, METH-induced hyperlocomotion, and METH self-administration in rats. However, these VMAT2 inhibitors lacked selectivity and tolerance developed to these pharmacologic effects after repeated administration, thereby limiting their potential as pharmacotherapeutics for METH use disorders. In the current study, analogs from a novel scaffold were found to selectively inhibit VMAT2 and were evaluated using neurochemical and behavioral pharmacological approaches. R- and S-3-(4-methoxyphenyl)-N-(1-phenylpropan-2-yl)propan-1-amine (GZ-11610 and GZ-11608, respectively) exhibited 94- to 3450-fold selectivity for VMAT2 over human-ether-a-go-go (hERG) channel, DAT, serotonin transporter, and nicotinic acetylcholine receptors. GZ-11608 competitively and concentration-dependently inhibited METH-evoked DA release via VMAT2. Also, GZ-11610 (56-300 mg/kg, oral) and GZ-11608 (300 mg/kg, oral; 10-30 mg/kg, s.c.) reduced METH-induced hyperlocomotor activity in METH-sensitized rats. Furthermore, GZ-11608 (1-30 mg/kg, s.c.) inhibited METH self-administration, cue- and METH-induced reinstatement in a dose-dependent manner, and 30 mg/kg (s.c.), 10 mg/kg (s.c.), and 17 mg/kg (s.c.) produced significant effect, respectively. Importantly, the GZ-11608-induced decrease in METH self-administration was not surmounted by increasing the amount of METH available. GZ-11608 did not substitute for METH and did not serve as a reinforcer in rats self-administering METH and drug naïve rats, respectively. Thus, these VMAT2 inhibitors incorporating a new scaffold are novel leads for new pharmacotherapeutics to treat METH use disorders. Substances with high abuse potential including opioids and cocaine elevate extracellular DA concentration in the nucleus accumbens, and this mechanism has long been considered to underly substance-induced reward. DA in the nucleus accumbens originates from DA neuron cell bodies located in the ventral tegmental area in the midbrain. Interestingly, M5 muscarinic acetylcholine receptors (mAChRs) are proteins that are highly expressed on ventral tegmental area DA neurons. Also, studies investigating M5 mAChRs knockout mice showed reduced responding for cocaine in cocaine self-administration and decreased time spent in cocaine-paired and morphine-paired place preference studies. Pharmacological inhibition of M5 mAChRs function via microinfusing mAChR antagonists exhibiting no selectivity among M1-M5 mAChRs subtypes into the ventral tegmental area where expression of M5 mAChRs are dominant, reduced morphine-induced hyperlocomotion and cocaine seeking behaviors in rats. These studies support therapeutic potential of M5 mAChRs selectivity antagonists in opioids and cocaine use disorders. Thus, in the current study, affinity of a series of pethidine and quinuclidinyl N-phenylcarbamate analogs for M5 mAChRs was evaluated using in vitro and ex vivo neuropharmacological assays. Among the pethidine analogs, compound 6a showed the highest binding affinity at M5 (Ki = 0.38 µM), but also high affinity at M1 and M3 mAChRs (0.67 and 0.37 µM, respectively). Among the quinuclidinyl N-phenylcarbamate analogs, compound 13c exhibited the highest affinity at M5 (Ki = 1.8 nM), but also high affinity at M1, M2, M3 and M4 mAChRs (Ki = 1.6, 13, 2.6, 2.2 nM, respectively). Also, 13c acted as an agonist of mAChRs on oxotremorine-induced DA release from rat striatal slices. In addition, compound 13b was found exhibiting the highest selectivity (17-fold) at M3 over M2 mAChRs, suggesting potential of 13b as a chronic obstructive pulmonary disease therapeutics. Taken together, these novel analogs serve as leads for further discovery of subtype-selective M5 mAChR antagonists that may have potential as therapeutics for substance use disorders, as well as for chronic obstructive pulmonary disease.

Research and development (R&D) innovation has become an important competitive advantage that is essential to the biopharma industry and critical to drug development (Bennani, 2012; Cuatrecasas, 2006; Douglas, et al., 2010; Garnier, 2008; Hine & Kapeleris, 2006; McKelvey, 2008; Pisano, 2006, 2010). Cultivating R&D innovation carries a distinct challenge of balancing between fostering conditions for creativity while at the same time managing for technical, scientific and operational efficiencies. However, the trend in the industry is toward using more process management techniques focused on formalization, standardization, control and efficiency in order to accelerate drug discovery efforts (Brown & Svenson, 1998; Cardinal, 2001; Cuatrecasas, 2006; Hine & Kaperleris, 2006; Johnstone et al., 2011; Paul et al., 2010; Scannell et al., 2012; Sewing et al., 2008; Ullman & Boutellier, 2008).

This study was designed to explore how early drug discovery scientists experience creativity in a highly coordinated and managed work environment. The research questions were as follows: How do scientists working in the early discovery phases of the R&D pipeline experience creativity; and how does a highly managed and coordinated work environment influence scientists’ experience of creativity? The basis for understanding their experiences was captured through detailed stories and reflections about their personal background, early influences and professional experiences as a scientist.

The sample included 10 early drug discovery scientists who work for either a pharmaceutical, biotech or bio-pharmaceutical company based in the United States of America (USA). A basic qualitative study was conducted with in-depth interviews as the primary method of data collection. Data were analyzed using multiple iterations of coding, describing and classifying to interpret what creativity means, how scientists experience creativity within their work environment and what factors influence this experience. Conclusions and implications about what creativity means, how scientists experience creativity and the various dynamics that shape this experience are presented in the following pages.

This dissertation presents a unique approach to human behavior analysis based on expert guidance and intervention through interactive construction and modification of behavior models. Our focus is to introduce the research area of behavior analysis, the challenges faced by this field, current approaches available, and present a new analysis approach: Interactive Relevance Search and Modeling (IRSM). More intelligent ways of conducting data analysis have been explored in recent years. Ma- chine learning and data mining systems that utilize pattern classification and discovery in non-textual data promise to bring new generations of powerful "crawlers" for knowledge discovery, e.g., face detection and crowd surveillance. Many aspects of data can be captured by such systems, e.g., temporal information, extractable visual information - color, contrast, shape, etc. However, these captured aspects may not uncover all salient information in the data or provide adequate models/patterns of phenomena of interest. This is a challenging problem for social scientists who are trying to identify high-level, conceptual patterns of human behavior from observational data (e.g., media streams). The presented research addresses how social scientists may derive patterns of human behavior captured in media streams. Currently, media streams are being segmented into sequences of events describing the actions captured in the streams, such as the interactions among humans. This segmentation creates a challenging data space to search characterized by non- numerical, temporal, descriptive data, e.g., Person A walks up to Person B at time T. This dissertation will present an approach that allows one to interactively search, identify, and discover temporal behavior patterns within such a data space. Therefore, this research addresses supporting exploration and discovery in behavior analysis through a formalized method of assisted exploration. The model evolution presented sup- ports the refining of the observer\'s behavior models into representations of their understanding. The benefit of the new approach is shown through experimentation on its identification accuracy and working with fellow researchers to verify the approach\'s legitimacy in analysis of their data.
Ph. D.

Foreign exchange market is the most active market in today’s global financial domains. While the consensus on several aspects of this market is fairly established, the informational efficiency in this market is still unsettled, particularly during unexpected interruptions and unusual or unstable periods. The financial crisis of 2008 is the most recent example of such a period. This dissertation focuses on the efficiency of the foreign exchange market during a unique, turbulent period using the six most actively traded currencies: the Australian dollar, Canadian dollar, Swiss franc, Euro, British pound, and Japanese yen. Considering nine months before the peak of the financial crisis to nine months thereafter, the entire sample is divided into three sub-samples: full-, non-crisis-, and crisis-periods. Both daily and minute-by-minute data are used. A variety of instruments are analyzed, including spot, forward, and exchange traded funds on the currencies. The methodologies that are employed range from standard econometric tests of efficiency to estimation of vector error correction models to identify price discovery, or leadership positions, in each of the currency markets. The findings indicate behavioral similarities and differences. The patterns of the volatility of the currencies are mixed: two-humped for the AUD, CAD, and EUR; W-shaped for the CHF; three-humped for the GBP, and flat U-shaped for the JPY. The daily results from several methodologies provide mixed evidence on market efficiency. Over the entire sample period, the estimated forward premium coefficients from the GARCH (1, 1) model are not significant for all currencies, while the null hypotheses of zero and one cointegrating vectors cannot be rejected for all currencies, except for the AUD. These findings are consistent with some of the previous studies, concluding that the efficiency tests in the foreign exchange market would depend on the methodology and the time period of the study. The high frequency data results show different degrees of price discovery between pair-wise instruments. Specifically, the spot exchange market shows a greater contribution to price discovery than the corresponding exchange traded funds. A possible explanation is the current size of the market and its increased transparency through the use of electronic trading.

Dairy farming as one part of agriculture has thousands of year’s history. The increasingdemands of dairy products and the rapid development of technology bring dairyfarming tremendous changes. Started by first hand milking, dairy farming goes throughvacuum bucket milking, pipeline milking, and now parlors milking. The automatic andtechnical milking system provided farmer with high-efficiency milking, effective herdmanagement and above all booming income.DeLaval Automatic Milking Rotary (AMRTM) is the world’s leading automatic milkingrotary system. It presents an ultimate combination of technology and machinerywhich brings dairy farming with significant benefits. AMRTM technical milking capacityis 90 cows per hour. However, constrained by farm management, cow’s condition andsystem configuration, the actual capacity is lower than technical value. In this thesis, anoptimization system is designed to analyze and improve AMRTM performance. The researchis focusing on cow behavior and AMRTM robot timeout. Through applying knowledgediscover from database (KDD), building machine learning cow behavior predictionsystem and developing modeling methods for system simulation, the optimizing solutionsare proposed and validated.
Mjölkproduktion är en del av vårt jordbruks tusenåriga historia. Med ökande krav påmejeriprodukter tillsammans med den snabba utvecklingen utav tekniken för det enormaförändringar i mjölkproduktionen. Mjölkproduktion började inledningsvis med handmjölkningsedan har mjölkproduktionsmetoder utvecklats genom olika tekniker och gettoss t.ex. vakuum mjölkning, rörledning mjölkning, fram till dagens mjölkningskarusell.Nu har det automatiska och tekniska mjölkningssystem försedd bönder med högeffektivmjölkning, effektiv djurhållningen och framför allt blomstrande inkomster.DeLaval Automatic Milking Rotary (AMRTM) är världens ledande automatiska roterandemjölkningssystemet. Den presenterar en ultimat kombination av teknik och maskinersom ger mjölkproduktionen betydande fördelar. DeLaval Automatic Milking Rotarytekniska mjölknings kapacitet är 90 kor per timme. Den begränsas utav jordbruksdrift,tillståndet hos kor och hantering av systemet. Det gör att den faktiska kapaciteten blirlägre än den tekniska. I denna avhandling undersöks hur ett optimeringssystem kan analyseraoch förbättra DeLaval Automatic Milking Rotary prestanda genom fokusering påkors beteenden och robot timeout. Genom att tillämpa kunskap från databas (KDD), skapamaskininlärande system som förutsäger kors beteenden samt utveckla modelleringsmetoderför systemsimulering, ges lösningsförslag av optimering samt validering.

Thesis (M. Eng.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 2011.
Cataloged from PDF version of thesis.
Includes bibliographical references (p. 121-126).
A content authoring bottleneck in AI, coupled with improving technology, has lead to increasing efforts in using large datasets to power Al systems directly. This idea is being used to create Al agents in video games, using logs of human-played games as the dataset. This new approach to AI brings its own challenges, particularly the need to annotate the datasets used. This thesis explores annotating the behavior in human-played games automatically, namely: how can we generate a list of events, with examples, describing the behavior in thousands of games. First dialogue is clustered semantically to simplify the game logs. Next, sequential pattern mining is used to find action-dialogue sequences that correspond to higher-level events. Finally, these sequences are grouped according to their event. The system can not yet replace human annotation, but the results are promising and can already help to significantly reduce the amount of human effort needed.
by Tynan S. Smith.
M.Eng.

The concepts of risk transfer and price discovery are well developed roles for futures markets. The interaction between traders in futures markets in the transferring and acceptance of price risk contributes to the discovery of price. Interaction of traders in the risk transfer and price discovery processes is examined in this dissertation. Data employed were for live cattle futures at the Chicago Mercantile Exchange developed from the confidential daily records of reporting trader positions maintained by the Commodity Futures Trading Commission. The analysis was for the period February 1983 through September 1987. The nearby futures contract price, volume, and open interest series supplement the daily trader position data base.
Ph. D.

The U.S. Department of Education has mandated that each U.S. state develop successful initiatives to help students navigate their educational experience. Yet in Alabama students neither advance academically nor in improved life skills development. It is unclear if school administrators in Alabama Schools have contextual best practices for strategic planning and implementation to support and improve the experiences of vulnerable K-12 students. The purpose of this descriptive case study was to explore how administrators of Alabama schools develop contextual best practices for strategic planning and implementation to support students. The conceptual framework was designed using collaboration theory, organizational learning theory, and appreciative inquiry. The overarching question addressed developing an understanding about how Alabama school administrators develop contextual best practices for strategic planning and implementation. Appreciative inquiry was used to facilitate a focus group and individual interviews with 15 participants. Data were analyzed using inductive analysis and bracketing. Thus, 4 themes were identified from the interviews and focus group. Most significant results were the identification of having a positive, engaging mobile environment and improving full community participation in the collaborative process. Contributions to positive social change may be experienced by developing community-based collaboration where all contribute to, and benefit from, co-create, collaborate, and structure a more balanced and feasible approach to successful implementation of strategic plans in an environment of financial constraints.

This research study is conducted to test Workman, Fiszman, Rindflesch and Nahl's information flow-serendipitous knowledge discovery (IF-SKD) model of information behavior, in a clinical care context. To date, there have been few attempts to model the serendipitous knowledge discovery of physicians. Due to the growth and complexity of the biomedical literature, as well as the increasingly specialized nature of medicine, there is a need for advanced systems that can quickly present information and assist physicians to discover new knowledge. The National Library of Medicine's (NLM) Lister Hill Center for Biocommunication's Semantic MEDLINE project is focused on identifying and visualizing semantic relationships in the biomedical literature to support knowledge discovery. This project led to the development of a new information discovery system, Spark. The aim of Spark is to promote serendipitous knowledge discovery by assisting users in maximizing the use of their conceptual short-term memory to iteratively search for, engage, clarify and evaluate information presented from the biomedical literature. Using Spark, this study analyzes the IF- SKD model by capturing and analyzing physician feedback. The McCay-Peet, Toms and Kelloway's Perception of Serendipity and Serendipitous Digital Environment (SDE) questionnaires are used. Results are evaluated to determine whether Spark contributes to physicians' serendipitous knowledge discovery and the ability of the IF-SKD ability to capture physicians' information behavior in a clinical setting.

The intraverbal is argued to be the most socially significant verbal operant and yet it is the least studied. Heal and Hanley (2011) suggest that different teaching strategies will lead to different rates of acquisition and child-preference with the tacting operant. This study continued this research into the realm of intraverbals, with focus on whether the embedded teaching strategy could be punishing on play or engaging in learning opportunities. The teaching strategies of discovery teaching, embedded prompting, and direct teaching were compared to see which strategy correlated with higher rates of acquisition and higher child preference. The study utilized a multi-element design by rapidly alternating teaching strategies while evaluating rate of acquisition and number of learning opportunities within the teaching strategies. Child preference was also demonstrated through card selection of associated teaching strategies in a concurrent chains agreement design. The teaching strategies differed in the amount of teacher directedness and taught intraverbal "Wh" questions. It was found through this study that embedded prompting did not punish play or the engagement in learning opportunities. The three participants preferred the three strategies differently and all participants were responding correctly the highest percentage of the time during the direct teaching contingencies by the end of the teaching sessions.

Aquesta tesi ofereix un examen en profunditat del comportament d’apreuament de les accions dipositades per part dels mercats emergents que, en gran part, s’han menystingut tot i el seu paper de dominació en l’àmbit del creuament de cotitzacions estrangeres. Les característiques de les accions dipositades fan que tinguin un valor idèntic al del seu stock subjacent i, per tant, esperen ser valorades de la mateixa manera. El fet de dur a terme una anàlisi detallada d’aquesta qüestió ha estat obstaculitzat fins ara per la falta de qualitat de les dades del dia a dia dels mercats emergents, unes informacions que faciliten l’anàlisi a temps real de la relació entre els preus de les accions dipositades i els seus stocks subjacents. Aquest estudi directe és necessari des del moment que aquests mercats tenen grans barreres comercials que possiblement distorsionen les relacions d’apreuament teòriques i emmascaren els veritables patrons d’apreuament. En el primer assaig s’examina el llarg funcionament fonamental de la relació econòmica que lliga els dos valors: la llei del preu únic. Proves recents demostren que, contràriament a allò que ocorre amb els valors de mercat, l’equitat de preus es trenca en els valors dels mercats emergents a causa de la presència de barreres comercials com ara els preus de comerç, les restriccions de la venda al detall i el control de capital. En el primer assaig es confirma la violació de la llarga carrera de la paritat en els preus de les accions dipositades egípcies, fet que està corroborat per proves de fortalesa dutes a terme durant diversos caps de setmana al mercat local i al mercat amfitrió, com també en els moviments del règim del tipus de canvi. El segon assaig se centra a identificar si existeixen oportunitats reals d’arbitratge quan es viola l’equilibri en la relació d’apreuaments subjacent. En l’anàlisi s’utilitza una única freqüència elevada durant dos anys de dades diàries de 16 valors egipcis i argentins per identificar si existeixen oportunitats d’arbitratge durant el període en què els dos valors estan comerciant i establir si les comercialitzacions d’arbitratge juguen algun paper en la convergència de preus. La metodologia emprada es basa en un procediment d’identificació del nou arbitratge que té en compte les despeses del comerç dinàmic i el volum d’aquest. S’ha establert una evidència de la presència d’un gran nombre d’oportunitats d’arbitratge a través de la mostra. S’ha vist que les oportunitats d’arbitratge persisteixen durant uns quants minuts i demanen més d’una comercialització per convergir en zones no arbitrades. A partir d’un algoritme de filtració s’extreuen les veritables comercialitzacions d’arbitratge de la comercialització d’arbitratge de les dades i s’estableix la importància del rol d’arbitratge en el moment de restablir els preus al seu valor fonamental i en el moment d’evitar que els preus vagin a la deriva lluny d’un preu implícit comú i eficient. El tercer assaig afegeix l’anàlisi de l’arbitratge i fa servir les mateixes dades diàries per examinar si és el mercat local o l’estranger el que juga un paper més dominant en l’apreuament diari dels valors de les cotitzacions creuades egípcies i argentines. Els resultats mostren que els dos mercats són importants per al procés de descobriment de preus, però que tant per a tots els valors egipcis com per a la major part dels argentins, el mercat local juga un paper més dominant. S’ha determinat que la localització de la descoberta del preu depèn de múltiples factors, entre els quals la liquiditat i el volum de comercialitzacions que cada mercat pot atraure són els més importants. El darrer assaig de la tesi va estar motivat pels resultats del tercer assaig i inspirat pels moviments de la Primavera Àrab de l’Orient Mitjà. L’aixecament a Egipte del 25 de gener va estar acompanyat pel tancament complet dels mercats d’stock durant dos mesos sencers. Aquest fet va crear una situació interessant en la qual els únics valors egipcis que podien comercialitzar eren aquells que tenien accions dipositades i que comercialitzaven al Regne Unit. Utilitzem aquest fet per examinar l’efecte d’un canvi en el marc legal en la localització de la descoberta del preu i per determinar que durant el període d’excepció durant el qual el mercat local va estar tancat, la localització de la descoberta del preu va canviar al mercat estranger, fent que aquest fos la localització dominant per a les activitats d’apreuament. Això proporciona dades de la naturalesa dinàmica del descobriment de preus de les accions dipositades.
Esta tesis ofrece un examen en profundidad del comportamiento de pricing de los recibos de depósito por parte de los mercados emergentes que, en gran parte, ha sido negligido a pesar de su papel dominante en el ámbito del cross-listing extranjero. Las características de los recibos de depósito hacen que sean títulos idénticos a su stock subyacente y, por tanto, se espera que sean valorados de la misma forma. El análisis detallado esta cuestión ha visto obstaculizado hasta ahora por la falta de datos de calidad intradía de los mercados emergentes, que facilite el análisis en tiempo real de la relación entre los precios de los recibos de depósito y su stock subyacente. Este análisis directo es necesario desde el momento que estos mercados tienen grandes barreras comerciales que posiblemente distorsionan las relaciones de pricing teóricas y enmascaran los verdaderos patrones de pricing. En el primer estudio, se examina la relación económica fundamental a largo plazo que relaciona los dos títulos: la ley del precio único. Pruebas recientes demuestran que, contrariamente a aquello que ocurre con los valores del mercado desarrollado, la paridad de precios se rompe en los valores de los mercados emergentes debido a la presencia de barreras comerciales como los precios comerciales, las restricciones de venta a corto y el control de capital. Este primer estudio confirma la violación de la paridad de precios de los recibos de depósito egipcios, hecho que está corroborado por los tests de fortaleza llevados a cabo durante varios fines de semana entre el mercado local y el de acogida, como también en los cambios de régimen del tipo de cambio. El segundo estudio se centra en identificar si existen oportunidades reales de arbitraje cuando se viola la relación de pricing de equilibrio subyacente. En el análisis, se usa una única serie de datos intradía de alta frecuencia durante dos años de 16 valores egipcios y argentinos para identificar si existen oportunidades de arbitraje durante el período en que los dos valores se están comerciando y establecer si las comercializaciones de arbitraje tienen algún papel en la convergencia de precios. La metodología usada se basa en un nuevo procedimiento de identificación del arbitraje que tiene en cuenta los costes comerciales dinámicos y los volúmenes. Se ha constatado que existe un gran número de oportunidades de arbitraje a lo largo de la muestra. Se ha visto que las oportunidades de arbitraje persisten durante unos cuantos minutos y requieren más de una comercialización para converger en zonas no arbitradas. A partir de un algoritmo de filtración, se extraen los intercambios reales de arbitraje de los intercambios de la serie de datos y se establece la importancia del rol de los árbitros para restablecer los precios a sus valores fundamentales y evitar que los precios se alejen de un precio implícito común y eficiente. El tercer estudio se basa en el análisis del arbitraje y utiliza la misma serie de datos intradía para examinar si es el mercado local o el extranjero el que tiene un papel más dominante en el pricing intradía de los valores de las cotizaciones cruzadas egipcias y argentinas. Los resultados muestran que los dos mercados son importantes para el proceso de revelación del precio, pero que para todos los valores egipcios y para la mayor parte de los argentinos el mercado local tiene un papel más dominante. Se ha observado que la localización de la revelación del precio depende de múltiples factores, entre los cuales la liquidez y el volumen de comercialización que cada mercado puede atraer son los más importantes. El último estudio de esta tesis fue motivado por los resultados del tercer estudio e inspirado por los movimientos de la Primavera Árabe del Oriente Medio. La revuelta del 25 de enero vino acompañado por el cierre total de los mercados bursátiles durante dos meses enteros. Este hecho creó un escenario interesante en el cual los únicos valores egipcios que se podían comercializar eran aquellos que tenían recibos de depósito y que comercializaban en el Reino Unido. Utilizamos este hecho para examinar el efecto de un cambio en el marco legal de la locación del descubrimiento del precio y para ver que durante el período de excepción durante el cual el mercado estuvo cerrado, la localización de la revelación del precio ha pasado al mercado emergente, haciendo que este fuera la localización dominante para las actividades de pricing. Esto demuestra la naturaleza dinámica del descubrimiento de precios de los recibos de depósito.
This thesis provides an in-depth examination of the pricing behavior of depository receipts from emerging markets which have been largely overlooked despite their dominating role in the foreign cross-listing arena. Characteristics of depository receipts make them identical securities to their underlying stock and therefore both are expected to be priced equally. A detailed analysis of the issue has been so far hampered by the lack of quality intraday data from emerging markets that facilitates a real time analysis of the relationship between the prices of the depository receipt and its underlying stock. This direct examination is required since those markets have large trading barriers that are hypothesized to distort the theoretical pricing relationship and mask true pricing patterns. The first essay examines the fundamental long run economic relationship that ties both securities: the law of one price. Recent evidence shows that contrary to developed market equities, price parity is broken in emerging market equities due to the presence of trading barriers such as trading costs, short selling restrictions and capital controls. The first essay confirms the violation of long run price parity in Egyptian depository receipts which is corroborated by robustness tests around the different weekends between the local and host market as well as around exchange rate regime shifts. The second essay focuses on identifying whether real arbitrage opportunities exist when the underlying equilibrium pricing relationship is violated. The analysis uses a unique two year high frequency intraday dataset from 16 Egyptian and Argentinean equities to identify whether arbitrage opportunities exist during the period when both securities are simultaneously trading and establish whether arbitrage trades play a role in price convergence. The methodology used relies on a novel arbitrage identification procedure that uses dynamic trading costs and volumes. Evidence of the presence of large number of arbitrage opportunities across the sample is established. Arbitrage opportunities are found to persist for several minutes and require more than one trade to converge to no-arbitrage zones. A filtering algorithm extracts real arbitrage trades from the arbitrage trades from the dataset and establishes the important role of arbitrageurs in restoring prices to their fundamental values and in keeping prices from drifting away from a common efficient implicit price. The third essay builds on the arbitrage analysis and uses the same intraday dataset to examine whether the local or foreign market plays a more dominant role in the intraday pricing of the Egyptian and Argentinean cross-listed securities. The results show that both markets are important for the price discovery process, but that for all of the Egyptian and most of the Argentinean securities, the local market plays a more dominant role. The location of price discovery is found to depend on several factors, most importantly the liquidity and trading volume that each market can attract. The final essay in the thesis was motivated by the results of the third essay and inspired by the Arab spring movements in the Middle East. The 25th of January uprising in Egypt was accompanied by a full stock market closure for a complete two months. This created an interesting setting in which the only Egyptian equities that were allowed to trade were those with depository receipts trading in the UK. We use this event to examine the effect of a change in the legal environment on the location of price discovery and find that during the interim period where the local market was closed, the location of price discovery has shifted to the foreign market making it the dominant location for pricing activity. This provides evidence of the dynamic nature of the price discovery of depository receipts.

Les progrès technologiques en systèmes de localisation (AIS, radar, GPS, RFID, etc.), de télétransmission (VHF, satellite, GSM, etc.), en systèmes embarqués et leur faible coût de production a permis leur déploiement à une large échelle. Énormément de données sur les déplacements d'objets sont produites par le biais de ces technologies et utilisées dans diverses applications de surveillance temps-réel comme la surveillance du trafic maritime. L'analyse a posteriori des données de déplacement de navires et d'événements à risques peut présenter des perspectives intéressantes pour la compréhension et l'aide à la modélisation des comportements à risques. Dans ce travail de thèse une méthodologie basée sur la fouille de données spatio-temporelle est proposée pour l'extraction de connaissances sur les comportements potentiellement à risques de navires. Un atelier d'aide à l'analyse de comportements de navires fondé sur cette méthodologie est aussi proposé
The advent of positioning system technologies (AIS, radar, GPS, RFID, etc.), remote transmission (VHF, satellite, GSM, etc.), technological advances in embedded systems and low cost production, has enabled their deployment on a large scale. A huge amount of moving objects data are collected through these technologies and used in various applications such as real time monitoring surveillance of maritime traffic. The post-hoc analysis of data from moving ships and risk events may present interesting opportunities for the understanding and modeling support of risky behaviors. In this work, we propose a methodology based on Spatio-Temporal Data Mining for the knowledge discovery about potentially risky behaviors of ships. Based on this methodology, a workshop to support the analysis of behavior of ships is also proposed

Within this thesis, we examined the possibilities of using an increasing amount ofpublicly available metadata about locations and peoples' activities in order to gainnew knowledge and develop new models of behavior and movement of people. Thepurpose of the research conducted for this thesis was to solve practical problems,such as: analyzing attractive tourist sites, defining the most frequent routes peopleare taking, defining main ways of transportation, and discovering behavioralpatterns in terms of defining strategies to suppress expansion of virus infections. Inthis thesis, a practical study was carried out on the basis of protected (aggregatedand anonymous) CDR (Caller Data Records) data and metadata of geo-referencedmultimedia content.
Предмет и циљ истраживања докторске дисертације представља евалуацијамогућности коришћења све веће количине јавно доступних података олокацији и кретању људи, како би се дошло до нових сазнања, развили новимодели понашања и кретања људи који се могу применити за решавањепрактичних проблема као што су: анализа атрактивних туристичких локација,откривање путања кретања људи и средстава транспорта које најчешћекористе, као и откривање важних параметара на основу којих се можеразвити стратегија за заштиту нације од инфективних болести итд. У раду је уту сврхе спроведена практична студија на бази заштићених (агрегираних ианонимизираних) ЦДР података и метаподатака гео-референцираногмултимедијалног садржаја. Приступ је заснован на примени техникавештачке интелигенције и истраживања података.
Predmet i cilj istraživanja doktorske disertacije predstavlja evaluacijamogućnosti korišćenja sve veće količine javno dostupnih podataka olokaciji i kretanju ljudi, kako bi se došlo do novih saznanja, razvili novimodeli ponašanja i kretanja ljudi koji se mogu primeniti za rešavanjepraktičnih problema kao što su: analiza atraktivnih turističkih lokacija,otkrivanje putanja kretanja ljudi i sredstava transporta koje najčešćekoriste, kao i otkrivanje važnih parametara na osnovu kojih se možerazviti strategija za zaštitu nacije od infektivnih bolesti itd. U radu je utu svrhe sprovedena praktična studija na bazi zaštićenih (agregiranih ianonimiziranih) CDR podataka i metapodataka geo-referenciranogmultimedijalnog sadržaja. Pristup je zasnovan na primeni tehnikaveštačke inteligencije i istraživanja podataka.

A causa de l’estreta relació entre la vida de les persones i determinades ubicacions geogràfiques, les dades històriques sobre trajectòries d’una persona contenen informació valuosa que es pot utilitzar per descobrir els seus estils de vida i hàbits. L’ús generalitzat de dispositius mòbils amb capacitat de localització ha impulsat la mineria de trajectòries (trajectory mining), la qual se centra en la manipulació, el processament i l’anàlisi de dades de trajectòries per facilitar l’extracció de coneixement a partir de l’històric de les trajectòries d’una persona. Basant-nos en aquesta anàlisi, fins i tot es pot arribar a predir quina serà la probable propera localització d’una persona. Amb aquestes tècniques, s’obre la porta a la millora dels actuals serveis basats en la ubicació i a l’aparició de nous models de negoci, basats en notificacions riques relacionades amb la predicció adequada de les futures ubicacions dels usuaris. Aquesta tesi tracta sobre la predicció de la ubicació i el descobriment de regions significatives a les zones de moviment de les persones. Proposa diversos models de predicció, basant-se en diferents tècniques d'aprenentatge automàtic (com ara les cadenes de Markov, les xarxes neuronals recurrents i les xarxes neuronals convolucionals), tot considerant diferents mètodes de representació d'entrada (embedding learning i one hot vector). A més, el model de predicció utilitza la attention technique (tècnica d’atenció), que té com a objectiu alinear els intervals de temps en les trajectòries de les persones que són rellevants per a una ubicació específica. La tesi també proposa un esquema de codificació temporal per capturar les característiques del comportament del moviment. Addicionalment, analitza l'impacte de l'aprenentatge de la representació espacial-temporal mitjançant l'avaluació de diferents arquitectures. Finalment, l’anàlisi de la trajectòria i la predicció de localització s’apliquen a la monitorització en temps real per a persones grans.
Debido a la estrecha relación entre la vida de las personas y determinadas ubicaciones geográficas, los datos históricos sobre trayectorias de una persona contienen información valiosa que se puede utilizar para descubrir sus estilos de vida y hábitos. El uso generalizado de dispositivos móviles con capacidad de localización ha impulsado la minería de trayectorias (trajectory mining), la cual se centra en la manipulación, el procesamiento y el análisis de datos de trayectorias para facilitar la extracción de conocimiento a partir de el histórico de las trayectorias de una persona. Basándonos en este análisis, incluso se puede llegar a predecir cuál será la probable próxima localización de una persona. Con estas técnicas, se abre la puerta a la mejora de los actuales servicios basados ​​en la ubicación y en la aparición de nuevos modelos de negocio, basados ​​en notificaciones ricas relacionadas con la predicción adecuada de las futuras ubicaciones de los usuarios. Esta tesis trata sobre la predicción de la ubicación y el descubrimiento de regiones significativas en las zonas de movimiento de las personas. Propone varios modelos de predicción, basándose en diferentes técnicas de aprendizaje automático (como las cadenas de Markov, las redes neuronales recurrentes y las redes neuronales convolucionales), considerando diferentes métodos de representación de entrada (embedding learning y one hot vector). Además, el modelo de predicción utiliza la attention technique (técnica de atención), que tiene como objetivo alinear los intervalos de tiempo en las trayectorias de las personas que son relevantes para una ubicación específica. La tesis también propone un esquema de codificación temporal para capturar las características del comportamiento del movimiento. Adicionalmente, analiza el impacto del aprendizaje de la representación espacial-temporal mediante la evaluación de diferentes arquitecturas. Finalmente, el análisis de la trayectoria y la predicción de localización se aplican a la monitorización en tiempo real para personas mayores.
Due to the relationship between people’s daily life and specific geographic locations, the historical trajectory data of a person contains lots of valuable information that can be used to discover their lifestyle and regularity. The generalisation in the use of mobile devices with location capabilities has fueled trajectory mining: the research area that focuses on manipulating, processing and analysing trajectory data to aid the extraction of higher level knowledge from the trajectory history of a user. Based on this analysis, even the person’s next probable location can be predicted. These techniques pave the way for the improvement of current location-based services and the rise of new business models, based on rich notifications related to the right prediction of users’ next location. This thesis addresses location prediction as well as the discovery of significant regions in person’s movement area. It proposes various models to predict the future state of people movement, based on different machine learning techniques (such as Markov Chains, Recurrent Neural Networks and Convolutional Neural Networks) and considering different input representation methods (embedding learning and one-hot vector). Moreover, the attention technique is used in the prediction model, aiming at aligning time intervals in people’s trajectories that are relevant to a specific location. Furthermore, the thesis proposes a time encoding scheme to capture movement behavior characteristics. In addition to that, it analyses the impact of Space-Time representation learning through evaluating different architectural configurations. Finally, trajectory analysis and location prediction is applied to real-time smartphone-based monitoring system for seniors.

Recentemente, várias tarefas tem sido delegadas a sistemas computacionais, principalmente quando sistemas computacionais são mais confiáveis ou quando as tarefas não são adequadas para seres humanos. O uso de extração de preferências ajuda a realizar a delegação, permitindo que mesmo pessoas leigas possam programar facilmente um sistema computacional com suas preferências. As preferências de uma pessoa são obtidas por meio de respostas para questões específicas, que são formuladas pelo próprio sistema computacional. A pessoa age como um usuário do sistema computacional, enquanto este é visto como um agente que age no lugar da pessoa. A estrutura e contexto das questões são apontadas como fonte de variações das respostas do usuário, e tais variações podem impossibilitar a factibilidade da extração de preferências. Uma forma de evitar tais variações é questionar um usuário sobre a sua preferência entre dois comportamentos observados por ele. A questão de avaliar relativamente comportamentos observados é mais simples e transparente ao usuário, diminuindo as possíveis variações, mas pode não ser fácil para o agente interpretar tais avaliações. Se existem divergências entre as percepções do agente e do usuário, o agente pode ficar impossibilitado de aprender as preferências do usuário. As avaliações são geradas com base nas percepções do usuário, mas tudo que um agente pode fazer é relacionar tais avaliações às suas próprias percepções. Um outro problema é que questões, que são expostas ao usuário por meio de comportamentos demonstrados, são agora restritas pela dinâmica do ambiente e um comportamento não pode ser escolhido arbitrariamente. O comportamento deve ser factível e uma política de ação deve ser executada no ambiente para que um comportamento seja demonstrado. Enquanto o primeiro problema influencia a inferência de como o usuário avalia comportamentos, o segundo problema influencia quão rápido e acurado o processo de aprendizado pode ser feito. Esta tese propõe o problema de Extração de Preferências com base em Comportamentos Observados utilizando o arcabouço de Processos Markovianos de Decisão, desenvolvendo propriedades teóricas em tal arcabouço que viabilizam computacionalmente tal problema. O problema de diferentes percepções é analisado e soluções restritas são desenvolvidas. O problema de demonstração de comportamentos é analisado utilizando formulação de questões com base em políticas estacionárias e replanejamento de políticas, sendo implementados algoritmos com ambas soluções para resolver a extração de preferências em um cenário sob condições restritas.
Recently, computer systems have been delegated to accomplish a variety of tasks, when the computer system can be more reliable or when the task is not suitable or not recommended for a human being. The use of preference elicitation in computational systems helps to improve such delegation, enabling lay people to program easily a computer system with their own preference. The preference of a person is elicited through his answers to specific questions, that the computer system formulates by itself. The person acts as an user of the computer system, whereas the computer system can be seen as an agent that acts in place of the person. The structure and context of the questions have been pointed as sources of variance regarding the users answers, and such variance can jeopardize the feasibility of preference elicitation. An attempt to avoid such variance is asking an user to choose between two behaviours that were observed by himself. Evaluating relatively observed behaviours turn questions more transparent and simpler for the user, decreasing the variance effect, but it might not be easier interpreting such evaluations. If divergences between agents and users perceptions occur, the agent may not be able to learn the users preference. Evaluations are generated regarding users perception, but all an agent can do is to relate such evaluation to his own perception. Another issue is that questions, which are exposed to the user through behaviours, are now constrained by the environment dynamics and a behaviour cannot be chosen arbitrarily, but the behaviour must be feasible and a policy must be executed in order to achieve a behaviour. Whereas the first issue influences the inference regarding users evaluation, the second problem influences how fast and accurate the learning process can be made. This thesis proposes the problem of Preference Elicitation under Evaluations over Observed Behaviours using the Markov Decision Process framework and theoretic properties in such framework are developed in order to turn such problem computationally feasible. The problem o different perceptions is analysed and constraint solutions are developed. The problem of demonstrating a behaviour is considered under the formulation of question based on stationary policies and non-stationary policies. Both type of questions was implemented and tested to solve the preference elicitation in a scenario with constraint conditions.

Ces dernière années, les systèmes de recommandation ont reçu beaucoup d'attention avec l'élaboration de nombreuses propositions qui tirent parti des nouvelles avancées dans les domaines du Machine Learning et du Deep Learning. Grâce à l'automatisation de la collecte des données des actions des utilisateurs tels que l'achat d'un objet, le visionnage d'un film ou le clic sur un article de presse, les systèmes de recommandation ont accès à de plus en plus d'information. Ces données sont des retours implicites des utilisateurs (appelé «~implicit feedback~» en anglais) et permettent de conserver l'ordre séquentiel des actions de l’utilisateur. C'est dans ce contexte qu'ont émergé les systèmes de recommandations qui prennent en compte l’aspect séquentiel des données. Le but de ces approches est de combiner les préférences des utilisateurs (le goût général de l’utilisateur) et la dynamique séquentielle (les tendances à court terme des actions de l'utilisateur) afin de prévoir la ou les prochaines actions d'un utilisateur. Dans cette thèse, nous étudions la recommandation séquentielle qui vise à prédire le prochain article/action de l'utilisateur à partir des retours implicites des utilisateurs. Notre principale contribution, REBUS, est un nouveau modèle dans lequel seuls les items sont projetés dans un espace euclidien d'une manière qui intègre et unifie les préférences de l'utilisateur et la dynamique séquentielle. Pour saisir la dynamique séquentielle, REBUS utilise des séquences fréquentes afin de capturer des chaînes de Markov d'ordre personnalisé. Nous avons mené une étude empirique approfondie et démontré que notre modèle surpasse les performances des différents modèles de l’état de l’art, en particulier sur des jeux de données éparses. Nous avons également intégré REBUS dans myCADservices, une plateforme collaborative de la société française Visiativ. Nous présentons notre retour d'expérience sur cette mise en production du fruit de nos travaux de recherche. Enfin, nous avons proposé une nouvelle approche pour expliquer les recommandations fournies aux utilisateurs. Le fait de pouvoir expliquer une recommandation permet de contribuer à accroître la confiance qu'un utilisateur peut avoir dans un système de recommandation. Notre approche est basée sur la découverte de sous-groupes pour fournir des explications interprétables d'une recommandation pour tous types de modèles qui utilisent comme données d’entrée les retours implicites des utilisateurs
Recommender systems have received a lot of attention over the past decades with the proposal of many models that take advantage of the most advanced models of Deep Learning and Machine Learning. With the automation of the collect of user actions such as purchasing of items, watching movies, clicking on hyperlinks, the data available for recommender systems is becoming more and more abundant. These data, called implicit feedback, keeps the sequential order of actions. It is in this context that sequence-aware recommender systems have emerged. Their goal is to combine user preference (long-term users' profiles) and sequential dynamics (short-term tendencies) in order to recommend next actions to a user. In this thesis, we investigate sequential recommendation that aims to predict the user's next item/action from implicit feedback. Our main contribution is REBUS, a new metric embedding model, where only items are projected to integrate and unify user preferences and sequential dynamics. To capture sequential dynamics, REBUS uses frequent sequences in order to provide personalized order Markov chains. We have carried out extensive experiments and demonstrate that our method outperforms state-of-the-art models, especially on sparse datasets. Moreover we share our experience on the implementation and the integration of REBUS in myCADservices, a collaborative platform of the French company Visiativ. We also propose methods to explain the recommendations provided by recommender systems in the research line of explainable AI that has received a lot of attention recently. Despite the ubiquity of recommender systems only few researchers have attempted to explain the recommendations according to user input. However, being able to explain a recommendation would help increase the confidence that a user can have in a recommendation system. Hence, we propose a method based on subgroup discovery that provides interpretable explanations of a recommendation for models that use implicit feedback

This project stems out of the need to improve the quality of life for both animals and humans, and facilitate the interaction between both environments into one habitat. The design will serve as a framework for the co-habitation and interaction of humans and animals in one habitat. At the core of this dissertation lies the concept of an animan habitat. This term embodies the concept of an intergrated habitat for all species. In arriving at the final design, a sequential thought process was applied. The logic behind this process will now be outlined (each corresponding section will be addressed in this document). Exploring the diversity of Architectural habitats and products has made it evident that involving Architects and Interior Architects into projects concerning animal space design can benefit wild animals and humans alike. Research into South African recreational nature spaces show the importance of the conservation of these existing spaces to different parties on local, national and international levels. The investigation of the importance of experiential nature spaces in Tshwane and the Tshwane CBDs (section 2.2.2) shows the importance and location of an establishment (the National Zoological Gardens of South Africa) with great human experiential, and animal conservation opportunities. The study of the contributions of global zoos to the world (section 2.3) confirms that any designs need to consider the principles and ethics followed by these zoos-conservation, recreation, education, experience, research, and community values. Behavioural enrichment (section 2.4) as a conservation contribution of zoos is an aspect that can be reinterpreted and incorporated into animal enclosures to enrich their environments, and further educate visitors. Studies about zoo evolutions (section 2.5) through the ages show how thoughts about captive environments are evolving. The subsequent study of exhibit design (section 2.6) makes clear the importance of considering the needs of the environment, animals, zoo occupants and visitors alike. Furthermore, research into design styles and illusions (section2.6) prove that designs (using whichever approach) should consider the wellbeing of animals before educating or entertaining humans. Design illusions could instead be used to change mans’ negative perceptions about zoos and other conserving environments. A study into design elements and principles (as studied by Ching and Miller) are currently used at the Zoo (section 2.7) to claim human and animal spaces. A variety of precedent investigations (section 3) make it clear that other institutions, zoos, reserves, bird parks, discovery centres, playgrounds, and an amphitheater, individuals (the work of Frei Otto) or companies (lightweight structure experts) offer products and techniques that could well suit animan space design. The result of the above is the cohabitation and respect for humans, animal and the environment in a bidirectional habitat. This forms the core of the animan concept and approach for the design of the Parrot Animan Precinct at the National Zoological Gardens of South Africa (Zoo). This development is a turnkey solution comprising of Site Selection and Study (section 4); Design Discourse (section 5); Technical Investigation (section 6) and Design Drawings (section 7).
Dissertation (MInt(Prof))--University of Pretoria, 2006.
Architecture
unrestricted

Esta pesquisa teve como objetivo principal desenvolver um modelo de comportamento de busca e recuperação da informação num ambiente ainda não explorado: os serviços de descoberta em rede. Para este fim, foi aplicada a metodologia da teoria fundamentada, o que permitiu conduzir, controlar e organizar a coleta de dados. Os dados formaram a base da teoria e a análise dos mesmos originou os conceitos que foram construídos. Os dados foram coletados durante a interação dos entrevistados com o serviço de descoberta em rede enquanto realizavam tarefas de busca e recuperação da informação. Os entrevistados foram selecionados aleatoriamente entre os usuários das bibliotecas da Universidad Complutense de Madrid. O modelo de comportamento de busca e recuperação da informação encontrado sugere duas etapas. O comportamento durante a busca relaciona-se com a maneira como os entrevistados reconhecem o problema, como exploram as opções de busca e como distinguem a informação para formular a expressão de busca. O comportamento de recuperação está relacionado à navegação e à diferenciação até a obtenção de uma lista de resultados considerados satisfatórios. A partir da lista de resultados é declarada a solução do problema por meio do reconhecimento da relevância, avaliação das fontes e seleção da informação. O referido modelo foi confrontado com modelos já referendados. Foram identificadas semelhanças com aqueles que pretendem entender o comportamento do usuário de forma mais ampla, levando em conta sua história, valores, contexto no qual está inserido e conhecimentos construídos a partir das experiências anteriores. Além disso, a pesquisa pretende colaborar com os aspectos metodológicos relacionados ao entendimento e a construção de uma teoria fundamentada. Conclui-se que é necessário proporcionar um maior conhecimento, aumentar a compreensão e fornecer um guia para a ação, com foco no usuário, para a adoção de serviços de descoberta de web em bibliotecas universitárias. Sugere aprimoramentos na sua customização e treinamentos para dotar os usuários de condições que lhes permitam extrair o máximo proveito possível das potencialidades deste tipo de ferramenta.
This research work aimed to develop a model of information seeking and retrieval behavior in an environment yet unexplored: the web discovery services. To this end, grounded theory methodologies were applied, which enabled the driving, tracking, and organizing of data collection. The data formed the basis of the theory and data analysis originated the concepts that were built. Data was collected during the interaction of interviewees with the web discovery service while performing tasks of information seeking and retrieval. The subjects were randomly selected from among the library users at the Universidad Complutense de Madrid. The information seeking and retrieval behavior model found suggests two steps. Behavior during searching is related to the way the respondents recognize the problem, exploit the searching options and distinguish information to formulate the expression. Behavior during information retrieval is related to navigation and differentiation to obtain a list of satisfactory results. From the list of results the solution of the problem is declared by recognizing the relevance, evaluation of sources and selection of information. The proposed model was confronted with models already established. Similarities to models that aim to understand user behavior more broadly, taking into account its history, values, context and knowledge constructed from previous experiences were identified. Besides, this research intends to collaborate with the methodological aspects related to the understanding and construction of a grounded theory. It concludes that is necessary to provide greater insight, enhance understanding and provide a guide to action, with a focus on the user, to the adoption of web discovery services in university libraries. It suggest customization and training to provide users with conditions that allow them to extract the maximum possible advantage of this type of tool.
Esta investigación tuvo como objetivo desarrollar un modelo de comportamiento para la búsqueda y recuperación de información en un entorno aún no explorado: las plataformas de descubrimiento. Para ello, se aplicó la metodología de la teoría fundamentada, lo que permitió dirigir, controlar y organizar la recopilación de datos. Los datos fueron la base de la teoría y del análisis se originaron los conceptos que fueron construidos. Los datos fueron recolectados durante la interacción de los participantes con la plataforma de descubrimiento mientras desempeñaban tareas de búsqueda y recuperación de información. Los encuestados fueron seleccionados al azar de entre los usuarios de las bibliotecas de la Universidad Complutense de Madrid. El modelo de comportamiento de búsqueda y de recuperación de la información encontrado sugiere dos etapas. El comportamiento durante la búsqueda se refiere a cómo los encuestados reconocen el problema, explotan las opciones de búsqueda y cómo distinguen la información para formular la expresión de búsqueda. El comportamiento durante la recuperación está relacionado con la navegación y la diferenciación para obtener una lista de resultados satisfactorios. En la lista de resultados se declara la solución del problema mediante el reconocimiento de la pertinencia, evaluación y selección de fuentes de información. El modelo se confronta a modelos ya aprobados. Se identificaron similitudes con los que tratan de comprender el comportamiento del usuario de manera más amplia, teniendo en cuenta su historia, valores, el contexto en el que se inserta y el conocimiento construido a partir de las experiencias anteriores. Por otra parte, la investigación tiene la intención de colaborar con los aspectos metodológicos relacionados con la comprensión y la construcción de una teoría fundamentada. Se concluye que es necesario proporcionar un mayor conocimiento, mejorar la comprensión y proporcionar un guía de acción, centrado en el usuario para la adopción de plataformas de descubrimiento en bibliotecas universitarias. Sugiere mejoras en su personalización y formación para dotar a los usuarios de las condiciones que les permitan sacar el máximo provecho posible del potencial de este tipo de herramienta.

Dopamine D3 receptors (DRD3) are implicated in relapse to drugs. The current study investigated the role of DRD3 in cue-induced reinstatement of nicotine-seeking in rats. Rats were trained to lever-press for intravenous infusions of nicotine, associated with the illumination of a cue-light, under a fixed-ratio schedule of reinforcement. Following extinction of the behaviour, where lever pressing had no consequences, reinstatement testing was performed by reintroduction of the cues after systemic or local administration (into discrete brain areas) of the DRD3 selective antagonist SB277011-A. Systemic antagonism of DRD3 significantly attenuated cue-induced reinstatement of nicotine-seeking. The same effect was observed upon infusions of SB277011-A into the basolateral amygdala or the lateral habenula, but not the nucleus accumbens. The current findings implicate DRD3 in cue-induced reinstatement of nicotine, delineate some of the neural substrates underlying this role and support a potential for using selective DRD3 antagonists for the prevention of relapse to smoking.

"Quorum sensing is a bacterial cell-cell communication mechanism mediated by the production and detection of small chemical molecules named autoinducers. This communication mechanism allows bacteria to regulate group behaviors, through synchronization of gene expression, in response to variation in cell density and species composition. The possibility of interfering with quorum sensing to control bacterial behaviors is an attractive alternative to the commonly used antibiotics, to which bacteria develop resistances. Autoinducer-2 (AI-2) mediated quorum sensing is especially interesting as it is involved in interspecies communication in complex biological niches, like the mammalian gut microbiota, where it was shown to affect the bacterial species composition.(...)"

博士
國立中央大學
財務金融學系
106
This study contains two essays on the price discovery and jump behaviors on VIX derivatives markets. The first essay investigates on price discovery across the S&P 500 index, S&P 500 index options, and VIX options markets; whereas the second essay provides a general form of multi-components option pricing model which includes multiple volatility, jump, co-jump, and leverage components, namely, Heterogeneous AutoRegerssive Gamma model for Realized Volatility with Leverage, Jumps, and Co-jumps (CoJJLHARG). First Essay: This paper investigates on price discovery across the S&P 500 index (SPX), SPX options, and VIX options markets by applying Hasbrouck’s (1995) information share and Yan and Zivot’s (2010) and Putniņš’s (2013) information leadership share methods. We estimated a time series regression model to integrate the price discovery into market characteristics. We also separated the data into two subsamples - one in the presence of crisis and the second in its absence - and examined the relationship between price discovery and market characteristics. In addition, this study provides a new angle to analyze whether the information is identical in the call and put options by market characteristics. Finally, this study contributes to literature since it indicates how informed traders in the option market are distributed across strike prices. Second Essay: This paper provides a general form of multi-components option pricing model which includes multiple volatility, jump, co-jump, and leverage components, namely, Heterogeneous AutoRegerssive Gamma model for Realized Volatility with Leverage, Jumps, and Co-jumps (CoJJLHARG). The model employs the high-frequency SPX and VIX data to filter the co-jump component. Moreover, we use this model to analyze the options pricing’s performance.

碩士
國立成功大學
國際經營管理研究所碩士在職專班
101
To find the internal service switching behavior in Taiwan IT industry. We conducted this article to connect and compare between the (un) satisfactory incidents. Critical incident technique was adopted for gathering over 67 e-service (traditional) incidents from service providers and receivers. Incident resolution was identified as the standard operation procedures (SOP). We got the meaningful findings that unsatisfactory services led to switch. However, the research participant looked forward adopting the non-switched reason and resolution. We grouped satisfactory solutions and SOP into subcategories to discover the behavioral status. Both internal service providers and receivers were proclaiming that they deserved to get more care (concern). The result was what MIS suggestion was not completely taken by receivers, vice and versa. The discoverable viewpoint was cross-department interviewees prefer to be treated beyond the 268 (67 multiplies 4) incidents, even though we had checked that the satisfactory solution might not necessarily lead to non-switch-behavior. In summary, it was MIS staffs obligation to treat equally on responses. The implication to the corporation was that to win staffs trust through communication, transform complaints into action, and put on-line suggestion into on-site practice. We proposed the satisfactory SOP to decrease switch, but the limitation was that we focused on eastern-context target. The beyond-incidents to the cross-industry collaboration / conflict were potential research issues.

碩士
中原大學
資訊管理研究所
107
Botnet threats continue to be a growing priority for organizations of all sizes in recent years. The designed malware of botnet is sophisticated and the corresponding communication behavior is inconspicuous. This paper introduces Visualize Intelligence and Temporal Analysis to network traffic as a framework to identify malware behavior hidden on the Internet. In this research, we condense traffic into a graphic and then utilize machine-learning algorithm to locate the behavior of beacon (BoB), which is a vital indication of auto-communication software. Since the malware within a compromised device will report to Command & Control (C&C) server periodically, the purpose of this research is to collect traffic flow and to discover the BoB by auto-learning algorithms, such as Artificial Neural Network. Our study confirms this framework model has exceptional performance and accuracy, as well as pinpointed the live beacon during investigation. Our study presents an analytical framework which takes into account the various beacons rate during the different time period. Extensive experimental result validates that framework has significant performance.

碩士
國立交通大學
管理學院碩士在職專班資訊管理組
97
Investment strategy is something that is important to the general public though at the same time is difficult to be formulated. Mutual fund, as one implementation of the investment strategy, may be a good fit to people without much time or willingness to track their investment portfolio on a frequent basis. The threshold for the mutual fund investment is relatively low. For example, to make single mutual fund investment locally may only cost ten thousands (10,000) NTD. And when people choose to investment their money on a monthly basis only three thousands (3,000) NTD may be enough. Regardless of the investor’s financial status, the mutual fund creates an opportunity for people to properly manage their fortunes by pooling their investment money, sharing the risks associated with the investment, and enjoying the profit together. However, any investment comes with the risks. Any investor including the mutual fund investor desires to maximize the profit of his/her investment portfolio while minimizing the loss. Any investor desires to purchase the mutual fund with a superior investment return ratio in the bull market. The investor may not want to be in possession of any mutual fund during the downturn of economy in order to avoid the loss, and even likes to reap certain profits by investment in futures. The present research picks and processes one hundred and eighty nine (189) domestic equity mutual funds and thirteen (13) macroeconomics indices to serve as inputs for self-organizing map neural networks so as to formulate a model for mutual fund market behavior and trend discovery. With such model in place, the present research further simulates investment transactions to verify its efficiency. Our verification shows the investment strategy formulated on the basis of the established model the overall investment return could be as high as one hundred and twenty two (122) percents of the investment while the Taiwan weighted stock index was down by twenty two (22) percents during the period from Jan., 2002 to Dec., 2008. That the average ratio of investment return stands at minus twenty-five (25) percents during the same period further suggests utilizing the model proposed by the present research for the mutual fund investment could outperform the random investment and stock transactions. The present research might also predict the trend of the stock market through buy/sell signals according to the established model, which further provides the investors with a valuable reference when it comes to selecting a target for the mutual fund investment.

碩士
國立臺灣師範大學
國際事務與全球戰略研究所
102
In this study, we modify the information share (IS) originally proposed by Hasbrouck, J.(1995) to discuss investors behavior which included domestic individual investor, domestic and foreign Institutional Investor. And this article also covered three markets to give example, such as Taiwan Stock Exchange Weighted Index (TWSE), Taiwan stock index option (TAIEX Options) and Taiwan stock index future(TAIEX Futures). The implied price of option here is using the Put-Call-Parity to replace the traditional B-S model. We check if there is any connection between those three kinds of investor to the Financial Instruments information share. The empirical results show that foreign Institutional Investor is taking an important role on the function of option price discovery. Hence, the evidence supports the viewpoints by Fisher(1966)、Cohan et al.(1986)、Lo and MacKinlay(1988)、Stoll and Whaley(1990). On the other hand, the domestic brokers are tending to without information advantage in this research period. And vice versa, the domestic individual investors have contributed to the price discovery function of Taiwan Stock Exchange Weighted Index market.

碩士
國立交通大學
管理學院碩士在職專班資訊管理組
97
In recent years, the rapidly fluctuations of macroeconomics to make business enterprises more difficult to get the right views about industry trend and product development trend and corporations to face financial crises depending on the wrong decision at the same time ,especially in the TFT-LCD field belonging to “Two Trillion Double Stars Facilitation Plan” in Taiwan. The major factor of business development is to completely catch the business cycle base on the prediction of world wide demand, expansion of facility, and product development trend. However, prior researches on those business cycle models were used to do estimate with statistic analysis, Expert Opinion or static financial statements to support their conclusions. This research base on the above factors expectation is totally masters the industry trend under the dynamic environment. It is such a good result for a lot of scholars to use an unsupervised learning, Self-Organizing Map (SOM) of a neural network, to analyze those financial crises and the trend of constitution judgments. Therefore, this study proposes to use Self-Organizing Map and hierarchical SOM on “Business Operation and Product Development Model”. Business Operation Model of this paper is to come out leading signal of industry trend by judging industry trend in static and dynamic points of view with using unsupervised clustering and visualization capability of SOM and following to have the moving the trajectory into two dimension girds and finally analyze the steps. And, the Product Development Trend Model is to rely on product related data to establish with one-step variable process, normalization, and the Hierarchical SOM. The present research of “Business Operation Model” plays a leading role to effectively identify the application of TFT-LCD information in the business cycle and. “Product Development Model” can propose a suggestion to make an investing decision.

博士
國立交通大學
資訊管理研究所
99
Global financial markets have been through an extreme downturn in the wake of crisis associated with sub-prime mortgages, causing many investors to suffer from an enormous loss of their investment. Despite the fact that investors generally will not be able to avoid any financial crisis of the similar scale from occurring again as is the case in high risk stock market which is based on the framework of speculations and business growth variance, one can reduce the risks probability with the aid of information technology. In the past two decades, financial derivatives have been gaining a significant popularity with the investors, employing a variety of financial derivatives as new tools and medium of investments for profiting.. Since an ETF (exchange-traded fund) is associated with characteristics of relatively lower transactional costs and risks along with relatively higher fluidity, more and more investors have been exploring and pursuing the ETF-related opportunities. Additionally, ETF-related financial products have been reaching beyond the country borders into different markets and to certain specific goods such as gold and minerals. The country-specific ETFs offer the profits of an international portfolio diversification at a lower cost with a lower tracking error in a more tax-efficient way, than passive open or closed-end country funds. The present research utilizes EWT at iShares and the corresponding researching period starts from 06/23/2000 to 06/16/2011. In the present research, data mining (DM) and neural network are both utilized for identifying the investment patterns associated with an exchange-traded fund (ETF). The present research further capitalizes on characteristics including low cost, tax and trading flexibility associated with an ETF to develop the investment strategy with high probability and a positive investment returns. Specifically, an approach proposed in the present study utilizes DM and BPNN. Experiment result shows that, “following the investment strategy prepared and utilized according to the present research will ensure better investment return than Random Walk methods”. As such, the present research work and model could provide the investors with appropriate guidance at the time of the decision-making in this constantly-changing financial market Investments.

碩士
國立交通大學
資訊管理研究所
102
There are many analytical methods applying to finance markets diversely. Technical analysis is often used by short or medium term investors, and MACD is one of the most popular indicators in technical analysis. There have been a lot of researches proving the effectiveness of MACD in the market. J.Peter Steidlmayer put forward market profile theory. The theory is that changes in market prices are caused by various types of market participants’ trading behavior. Market does not follow random walk, but it fluctuates with the behavior patterns and rules. In market profile theory, participants are divided into long-term and short-term traders, market activities into initiative and responsive activity, trading types into buying and selling. There are at least eight behaviors of market participants. However, lack of researches were conducted on the market profile theory in the financial field. Hence, this research combine the MACD crossover theory and the behavior of market profile. The research is conducted in the purpose of trying to explore the relationships between both theories, and to build an effective model based on the results of the statistical analysis. By combining these two, we expect that the model is more effective than the model merely using the MACD indicator. Results of this research shows that the model combining MACD and market profile has better performance than the model that only use MACD indicator. In conclusion, it can effectively help investors grasp market trends.

碩士
國立中央大學
資訊工程學系
107
With the rapid development of information technology and the popularity of smart devices, users' demand for instant processing of network services and diversified services has also increased significantly, making the architecture of traditional network services unable to meet the rapidly changing network architecture of emerging services Demand. Software-defined Networking (SDN) and Network Function Virtualization (NFV) have therefore been proposed to transform complex network architectures into virtual and programmable architectures to reduce network complexity, bringing about major changes to the traditional network architecture. SDN controller use OpenFlow Discovery Protocol (OFDP), which detects the links between the OpenFlow switches by generating Link Layer Discovery Protocol (LLDP) packets, to collect comprehensive network topology status for the routing and switching of packets. However, OFDP is not a completely secure protocol and can be used by attackers to perform topology discovery injection attack, topology discovery man-in-the-middle attack and topology discovery flood attack, thereby confusing the network topology.

Tese de doutoramento em Tecnologias e Sistemas de Informação (área de especialização em Engenharia e Gestão de Sistemas de Informação)
Nesta tese apresenta-se uma ferramenta para gestão, integração e consolidação do Conhecimento sobre o Comportamento dos Clientes (CCC) obtido a partir da actuação de ferramentas de data mining sobre bases de dados transaccionais de organizações. As ferramentas de data mining permitem automatizar a detecção de padrões de comportamento dos clientes de uma organização a partir das bases de dados transaccionais, num processo designado por Descoberta do Conhecimento em Bases de Dados (DCBD). Estes padrões podem ser transmitidos aos agentes organizacionais e utilizados em campanhas de marketing e outras actividades no contexto da organização. No entanto, este conhecimento sobre o comportamento dos clientes não é, normalmente, objecto de qualquer tratamento que permita a análise das razões para o seu aparecimento ou da sua evolução bem como a consolidação com outro conhecimento sobre o CCC já existente. Há, pois, neste processo, uma situação que consideramos que pode ser melhorada com a introdução dum novo conceito - a Gestão do CCC – o qual conduz a uma nova actividade organizacional – Zelar pelo CCC. A Gestão do Conhecimento sobre o Comportamento dos Clientes é entendida como a confrontação deste conhecimento com outro conhecimento já existente na organização, resolvendo potenciais conflitos, actualizando-o e acrescentando explicações pertinentes para a evolução temporal verificada. As principais contribuições deste trabalho centram-se: - na apresentação do conceito “a Gestão do Conhecimento sobre o Comportamento dos Clientes” que conduz a uma nova tarefa organizacional “Zelar pelo Conhecimento sobre o Comportamento dos Clientes”; - no estabelecimento de uma forma de estrutura do CCC e seu registo; - na concepção e exploração dum sistema de conhecimento para apoio à gestão do conhecimento do comportamento dos clientes - o sistema ZACCAR (Zelar pela Aquisição do Conhecimento dos Clientes, sua Actualização e Registo) - cujo objectivo principal é permitir a viabilidade da nova tarefa através da: recolha e uniformização dos padrões de comportamento obtidos com uma ferramenta de data mining; confrontação desses padrões com o conhecimento já existente acerca do comportamento dos clientes, actualizando-o; validação e documentação, pelo gestor do conhecimento organizacional, do conhecimento já actualizado; integração do conhecimento depois de actualizado e completado, numa base de conhecimento que fará parte integrante do conhecimento organizacional; - no processo de consolidação do conhecimento descoberto em bases de dados, resolvendo problemas de interpretação, integração e conflitos. Na prossecução dos objectivos que estiveram presentes na elaboração deste trabalho, foi feita uma análise pormenorizada da prática de CRM (Customer Relationship Management) e sua relação com o conhecimento organizacional bem como do CCC com ênfase no tratamento que é dado a este conhecimento. O sistema ZACCAR pode-se considerar um sistema inovador uma vez que permite às organizações dispor de uma base de conhecimento do CCC, actualizada duma forma semi-automática onde está, ainda, registada uma evolução dos padrões de comportamento dos clientes e que faz parte integrante do conhecimento organizacional. Um protótipo do ZACCAR foi desenvolvido, recorrendo a tecnologia existente; para demonstrar a sua exequibilidade, foram efectuados dois estudos de casos os quais demonstram que o sistema possui potencialidades interessantes que se poderão tornar muito úteis em qualquer empresa onde o sistema seja implantado quer como sistema independente quer como integrado noutros sistemas empresariais de maior abrangência.
In this thesis it is presented a tool to take care of the Customers’ Behaviour Knowledge (CBK) obtained when a data mining tool acts in the organisational databases to manage and integrate it in the organisational knowledge, through a consolidation process with the existing knowledge. Data mining tools automate the detection of customers'behaviour patterns from the organisational databases in a process called Knowledge Discovery in Databases (KDD). These patterns may be transmitted to organisational agents and used in marketing campaigns and other activities in the organisation. However, the CBK is not usually treated to allow the analysis why it exists or how it evolves as well its consolidation with other existing CBK. So, we consider that, in this process, there is a situation that can be optimized through the introducing of a new concept - the management of CBK - conducting to a new organisational activity - to take care of the CBK. The management of CBK is intended as the confrontation of this knowledge with other existing knowledge, resolving potential conflicts, updating it and adding pertinent explanations to the temporal evolution of the customers'behaviour patterns. The most important contributions of this work are: - the presentation of the concept "The management of the Customers’ Behaviour Knowledge" that allows a new organisational task: "To take care of the Customers’ Behaviour Knowledge"; - the creation and exploration of a knowledge system to help the management of the CBK - the ZACCAR system - whose main objective is to permit the viability of the new task that is got by: the collecting and uniformization of the behaviour patterns obtained with a data mining tool; the confrontation of these patterns with existing CBK, updating it; the validation and documentation, by the manager of the organisational knowledge, of the knowledge after to be updated; the integration of the updated knowledge in a knowledge base that will be an integrant part of the organisational knowledge; - in the consolidation process of the knowledge discovered in databases, resolving interpreting and integration problems as possible conflicts. Attending the objectives considered in this work, it was made a detailed analysis of the practice of CRM (Customer Relationship Management) and its relation with the organisational knowledge as well of the CBK with emphasis in the treatment given to this knowledge ZACCAR can be considered an innovating system as, with it, the organisations can have a knowledge base of the CBK, updated in a semi-automatic process where it can be yet, stored the evolution of the customers'behaviour patterns and turned as an integrant part of the organisational knowledge. It was developed a prototype of ZACCAR, using existing technology; to prove its feasibility it was conducted two case studies; these cases showed that the system has good potentialities that will be very useful in an enterprise where the system can be implemented either as independent system or integrated in other organisational systems with a greater covering.
Projecto parcialmente financiado por uma bolsa do PRODEP II, medida 5, acção 5.2, concurso nº1/96, Doutoramentos.

碩士
國立臺灣師範大學
社會教育學系
96
“The Discovery Center of Taipei” provides various types of exhibitions for visitors to interact with to present the history and life facts of Taipei. In Taiwan, the majority of master’s theses related to the audience behavior took the scientific type of museum as a research area; only few researches focuses on social historical type of museums. The purpose of this study is to explore the exhibitions and the adult visitors’ behavior in “The Discovery Center of Taipei” in order to improve the exhibitions of social history museums. This study took the “Dialogue with Time Hall” at The Discovery Center of Taipei as the research area between March to May 2008. This study used content analysis in the exhibition aspect and also used nonparticipant observation and semi-structured interview to get the attracting power and holding power of adult visitors for each exhibition. Adult visitors’ behavior is understood, too. The findings of this study are: (1) There are some relationships between the attracting power and holding power of exhibitions. (2) Different exhibitions affect attracting power and holding power. The exhibitions which adult visitors can interact with or provide highly sense stimulate can get higher attracting power and holding power. (3) Different type of exhibit strategies attract different kind of adult visitors. (4) Adult visitors’ behavior is mostly in the visual behavior, and is influenced by personal, social and physical contexts. According to the findings, this study makes the following suggestions： (1) At the suggestions to the practice: to satisfy the need of adult visitors; to value adult visitors’ learning characters in museums; to improve the cues to hint adult visitors to interact with the exhibitions; to maintain the operation of the international exhibitions. (2) At the suggestions to the future researches: to gain quantification data to explore the adult visitors’ demographic characteristics and preferences and to broaden the research scope to social history museums.

碩士
國立交通大學
管理學院碩士在職專班資訊管理組
98
Because the rapid change of information transmission, Taiwan stock market is also easy to influence by unexpected events. Many investors’ property changes in the stock market overnight sometimes shrink substantially. Therefore, it is urgent that the stock market investors seeking low risk investment opportunities to reduce the unpredictable risk. In order to prevent the dramatic overnight losses, the long-term investors need to build a hedging model to save their own property. In this study, we apply the theory of artificial intelligence in the field of back-propagation neural network to clustering the historical data of the behavior of opening patterns after 15 minutes in Taiwan weighted index price by the time 09:05, 09:10, 09:15 of the closing price. Produce eight types of groups, then each group of data entry to the back-propagation neural networks to predict relative to the same day's closing price of the Taiwan stock price index, and tests the investors in Taiwan's futures index as trading partners.The experimental result confirmed that after the experimental model through the combination of clustering propagation neural network to predict the exact rate was significantly better than the control group which only using back propagation neural network and the random walk model. In addition, the model with M2 (up, up, down), M7 (down, down, up) of the investment transaction accuracy and the profitability are the best profit performance model. Therefore, these experiments assisted by clustering has better grasp of the changes in the environment to make dynamic learning. Thus provide investors with more specific transactions information to assist decision-makers to make the right choice.