To see the other types of publications on this topic, follow the link: CLASSIFICATION OF RESEARCH.
Dissertations / Theses on the topic 'CLASSIFICATION OF RESEARCH'
Create a spot-on reference in APA, MLA, Chicago, Harvard, and other styles
Consult the top 50 dissertations / theses for your research on the topic 'CLASSIFICATION OF RESEARCH.'
Next to every source in the list of references, there is an 'Add to bibliography' button. Press on it, and we will generate automatically the bibliographic reference to the chosen work in the citation style you need: APA, MLA, Harvard, Chicago, Vancouver, etc.
You can also download the full text of the academic publication as pdf and read online its abstract whenever available in the metadata.
Browse dissertations / theses on a wide variety of disciplines and organise your bibliography correctly.
1
Francis, Paul John. "The classification of quasar spectra." Thesis, University of Cambridge, 1991. http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.239185.
Full text
2
Wei, Zhihua. "The research on chinese text multi-label classification." Thesis, Lyon 2, 2010. http://www.theses.fr/2010LYO20025/document.
Full textAbstract:
Text Classification (TC) which is an important field in information technology has many valuable applications. When facing the sea of information resources, the objects of TC are more complicated and diversity. The researches in pursuit of effective and practical TC technology are fairly challenging. More and more researchers regard that multi-label TC is more suited for many applications. This thesis analyses the difficulties and problems in multi-label TC and Chinese text representation based on a mass of algorithms for single-label TC and multi-label TC. Aiming at high dimensionality in feature space, sparse distribution in text representation and poor performance of multi-label classifier, this thesis will bring forward corresponding algorithms from different angles.Focusing on the problem of dimensionality “disaster” when Chinese texts are represented by using n-grams, two-step feature selection algorithm is constructed. The method combines filtering rare features within class and selecting discriminative features across classes. Moreover, the proper value of “n”, the strategy of feature weight and the correlation among features are discussed based on variety of experiments. Some useful conclusions are contributed to the research of n-gram representation in Chinese texts.In a view of the disadvantage in Latent Dirichlet Allocation (LDA) model, that is, arbitrarily revising the variable in smooth process, a new strategy for smoothing based on Tolerance Rough Set (TRS) is put forward. It constructs tolerant class in global vocabulary database firstly and then assigns value for out-of-vocabulary (oov) word in each class according to tolerant class.In order to improve performance of multi-label classifier and degrade computing complexity, a new TC method based on LDA model is applied for Chinese text representation. It extracts topics statistically from texts and then texts are represented by using the topic vector. It shows competitive performance both in English and in Chinese corpus.To enhance the performance of classifiers in multi-label TC, a compound classification framework is raised. It partitions the text space by computing the upper approximation and lower approximation. This algorithm decomposes a multi-label TC problem into several single-label TCs and several multi-label TCs which have less labels than original problem. That is, an unknown text should be classified by single-label classifier when it is partitioned into lower approximation space of some class. Otherwise, it should be classified by corresponding multi-label classifier.An application system TJ-MLWC (Tongji Multi-label Web Classifier) was designed. It could call the result from Search Engines directly and classify these results real-time using improved Naïve Bayes classifier. This makes the browse process more conveniently for users. Users could locate the texts interested immediately according to the class information given by TJ-MLWCLa thèse est centrée sur la Classification de texte, domaine en pleine expansion, avec de nombreuses applications actuelles et potentielles. Les apports principaux de la thèse portent sur deux points : Les spécificités du codage et du traitement automatique de la langue chinoise : mots pouvant être composés de un, deux ou trois caractères ; absence de séparation typographique entre les mots ; grand nombre d’ordres possibles entre les mots d’une phrase ; tout ceci aboutissant à des problèmes difficiles d’ambiguïté. La solution du codage en «n-grams »(suite de n=1, ou 2 ou 3 caractères) est particulièrement adaptée à la langue chinoise, car elle est rapide et ne nécessite pas les étapes préalables de reconnaissance des mots à l’aide d’un dictionnaire, ni leur séparation. La classification multi-labels, c'est-à-dire quand chaque individus peut être affecté à une ou plusieurs classes. Dans le cas des textes, on cherche des classes qui correspondent à des thèmes (topics) ; un même texte pouvant être rattaché à un ou plusieurs thème. Cette approche multilabel est plus générale : un même patient peut être atteint de plusieurs pathologies ; une même entreprise peut être active dans plusieurs secteurs industriels ou de services. La thèse analyse ces problèmes et tente de leur apporter des solutions, d’abord pour les classifieurs unilabels, puis multi-labels. Parmi les difficultés, la définition des variables caractérisant les textes, leur grand nombre, le traitement des tableaux creux (beaucoup de zéros dans la matrice croisant les textes et les descripteurs), et les performances relativement mauvaises des classifieurs multi-classes habituels
文本分类是信息科学中一个重要而且富有实际应用价值的研究领域。随着文本分类处理内容日趋复杂化和多元化，分类目标也逐渐多样化，研究有效的、切合实际应用需求的文本分类技术成为一个很有挑战性的任务，对多标签分类的研究应运而生。本文在对大量的单标签和多标签文本分类算法进行分析和研究的基础上，针对文本表示中特征高维问题、数据稀疏问题和多标签分类中分类复杂度高而精度低的问题，从不同的角度尝试运用粗糙集理论加以解决，提出了相应的算法，主要包括：针对n-gram作为中文文本特征时带来的维数灾难问题，提出了两步特征选择的方法，即去除类内稀有特征和类间特征选择相结合的方法，并就n-gram作为特征时的n值选取、特征权重的选择和特征相关性等问题在大规模中文语料库上进行了大量的实验，得出一些有用的结论。针对文本分类中运用高维特征表示文本带来的分类效率低，开销大等问题，提出了基于LDA模型的多标签文本分类算法，利用LDA模型提取的主题作为文本特征，构建高效的分类器。在PT3多标签分类转换方法下，该分类算法在中英文数据集上都表现出很好的效果，与目前公认最好的多标签分类方法效果相当。针对LDA模型现有平滑策略的随意性和武断性的缺点，提出了基于容差粗糙集的LDA语言模型平滑策略。该平滑策略首先在全局词表上构造词的容差类，再根据容差类中词的频率为每类文档的未登录词赋予平滑值。在中英文、平衡和不平衡语料库上的大量实验都表明该平滑方法显著提高了LDA模型的分类性能，在不平衡语料库上的提高尤其明显。针对多标签分类中分类复杂度高而精度低的问题，提出了一种基于可变精度粗糙集的复合多标签文本分类框架，该框架通过可变精度粗糙集方法划分文本特征空间，进而将多标签分类问题分解为若干个两类单标签分类问题和若干个标签数减少了的多标签分类问题。即，当一篇未知文本被划分到某一类文本的下近似区域时，可以直接用简单的单标签文本分类器判断其类别；当未知文本被划分在边界域时，则采用相应区域的多标签分类器进行分类。实验表明，这种分类框架下，分类的精确度和算法效率都有较大的提高。本文还设计和实现了一个基于多标签分类的网页搜索结果可视化系统（MLWC），该系统能够直接调用搜索引擎返回的搜索结果，并采用改进的Naïve Bayes多标签分类算法实现实时的搜索结果分类，使用户可以快速地定位搜索结果中感兴趣的文本。
3
Mayo, Robert William. "An evaluation of social grade as a classification scheme." Thesis, City University London, 2000. http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.312902.
Full text
4
Dunlap, James 1963. "Classification and analysis of longwall delays." Thesis, Virginia Tech, 1990. http://hdl.handle.net/10919/42403.
Full text
5
Egan, Shaun Peter. "A framework for high speed lexical classification of malicious URLs." Thesis, Rhodes University, 2014. http://hdl.handle.net/10962/d1011933.
Full textAbstract:
Phishing attacks employ social engineering to target end-users, with the goal of stealing identifying or sensitive information. This information is used in activities such as identity theft or financial fraud. During a phishing campaign, attackers distribute URLs which; along with false information, point to fraudulent resources in an attempt to deceive users into requesting the resource. These URLs are made obscure through the use of several techniques which make automated detection difficult. Current methods used to detect malicious URLs face multiple problems which attackers use to their advantage. These problems include: the time required to react to new attacks; shifts in trends in URL obfuscation and usability problems caused by the latency incurred by the lookups required by these approaches. A new method of identifying malicious URLs using Artificial Neural Networks (ANNs) has been shown to be effective by several authors. The simple method of classification performed by ANNs result in very high classification speeds with little impact on usability. Samples used for the training, validation and testing of these ANNs are gathered from Phishtank and Open Directory. Words selected from the different sections of the samples are used to create a `Bag-of-Words (BOW)' which is used as a binary input vector indicating the presence of a word for a given sample. Twenty additional features which measure lexical attributes of the sample are used to increase classification accuracy. A framework that is capable of generating these classifiers in an automated fashion is implemented. These classifiers are automatically stored on a remote update distribution service which has been built to supply updates to classifier implementations. An example browser plugin is created and uses ANNs provided by this service. It is both capable of classifying URLs requested by a user in real time and is able to block these requests. The framework is tested in terms of training time and classification accuracy. Classification speed and the effectiveness of compression algorithms on the data required to distribute updates is tested. It is concluded that it is possible to generate these ANNs in a frequent fashion, and in a method that is small enough to distribute easily. It is also shown that classifications are made at high-speed with high-accuracy, resulting in little impact on usability.
6
Qi, Wang. "Studies in the Dynamics of Science : Exploring emergence, classification, and interdisciplinarity." Doctoral thesis, KTH, Industriell ekonomi och organisation (Inst.), 2016. http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-184724.
Full textAbstract:
The dynamic nature of science is embodied in the growth of knowledge in magnitude and the transformation of knowledge in structure. More specifically, the growth in magnitude is indicated by a sharp increase in the number of scientific publications in recent decades. The transformation of knowledge occurs as the boundaries of scientific disciplines become increasingly less distinct, resulting in a complicated situation wherein disciplines and interdisciplinary research topics coexist and co-evolve. Knowledge production in such a context creates challenges for the measurement of science.This thesisaims to develop more flexible bibliometric methodologies in order to address some of the challenges to measuring science effectively. To be specific, this thesis1) proposes a new approach for identifying emerging research topics; 2) measuresthe interdisciplinarity of research topics; 3) explores the accuracy of the journal classification systems of the Web of Science and Scopus; 4) examines the role of cognitive distance in grant decisions; and 5) investigates the effect of cognitive distance between collaborators on their research output. The data used in this thesisaremainly from the in-house Web of Science and Scopus databases of the Centre for Science and Technology Studies (CWTS) at Leiden University. Quantitativeanalyses, in particular bibliometric analyses,are the main research methodologies employed in this thesis. Furthermore, this thesis primarily offers methodological contributions, proposing a series of approaches designed to tackle the challenges created by the dynamics of science. While the major contribution of this dissertation lies in the improvement of certain bibliometric approaches, it also enhances the understanding of the current system of science. In particular, the approaches and research findings presented here have implications for various stakeholders, including publishing organizations, bibliographic database producers, research policy makers, and research funding agencies. Indeed, these approaches could be built into a software tool and thereby be made available to researchers beyond the field of bibliometric studies.QC 20160406
7
Vaidya, Priyanka S. "Artificial Intelligence Approach to Breast Cancer Classification." University of Akron / OhioLINK, 2009. http://rave.ohiolink.edu/etdc/view?acc_num=akron1240957599.
Full text
8
Monaghan, Mark Peter. "The turmoil of evidence : research utilisation in UK drug classification." Thesis, University of Leeds, 2008. http://etheses.whiterose.ac.uk/701/.
Full textAbstract:
This thesis investigates the thorny relationship between evidence utilisation and policy making in a heavily politicised policy area. Expectations for the conflux of researcha nd policy formulation have been consolidated in the last decade under the banner of 'evidence-based policy'. In recent times, the debateo ver the nature and utility of evidence-based policy has become much more sophisticated. No longer can the connection between evidence utilisation and policy formulation be conceived in terms of evidence shaping policy outcomes or, conversely, policy being evidence free, where evidence has no impact. Such conceptualisations persist, however, in heavily politicised policy areas, where there is intense media scrutiny of decision making, a lack of consensus on its direction, prolonged conflict between competing interest and stakeholder groups and a permeating sense of crisis. These tend to relate to more 'macro' policy areas, not usually the remit of evidence-based policy-making and evaluative research. Using recent and ongoing developments in UK drug classification policy as a case-study, an explanatory framework of the complex role and nature of evidence in heavily politicised policy areas is developed. Central to this, is the use of a methodological approach that can account for the role of conflict in the policy process. A modified version of the Advocacy Coalition Framework is employed to this end. This, in turn, allows for a range of data-collection methods to be used, including observation and documentary analysis of Parliamentary Select Committee hearings alongside qualitative interviews with a wide-range of key policy actors involved in the decision-making process. From this a nuanced account of the evidence and policy relationship in such contexts is ascertained,which departs from the more established models explaining the evidence and policy nexus. Traditionally, such explanations have been conceived as models of research utilisation. In this research it is suggested that these do not translate effectively as models of evidence-based policy-making. This is because they are beset with some, or all, of the following problems: a) they focus more on 'research' rather than the broader concept of 'evidence'; b) they operate with a static view of the policy process where there is a direct connection between research and policy; c) they restrict the role of evidence to one of policy outcomes, rather than viewing the role of evidence in the process of decision-making; d) they assume that research is the defining influence on the decision-making processe; e) they operate at a high level of abstraction, offering little account of how research is selected for use in decision making. Consequently, a newer addition to the literature is developed, which, it is claimed, avoids these shortcomings.
9
Dutta, Bidyarthi, Krishnapada Majumder, and B. K. Sen. "Classification of Keywords Extracted from Research Articles Published in Science Journals." National Institute of Science Communication and Information Resources, 2008. http://hdl.handle.net/10150/105938.
Full textAbstract:
This paper is based on an analytical study of 335 keywords extracted from titles and abstracts of 70 research articles, taking ten from each year starting from 2000 to 2006, in decreasing order of relevance, on the subject Fermi Liquid, which is a specific subject under the broad area of Condensed Matter Physics. The research articles have been collected from the bibliographic database of INSPEC. The keywords are indexed to critically examine its physical structure that is composed of three fundamental kernels, viz. keyphrase, modulator and qualifier. The keyphrase reflects the central concept, which is usually post-coordinated by the modulator to amend the central concept in accordance with the relevant context. The qualifier comes after the modulator to describe the particular state of the central concept and/or amended concept. The keywords are further classified in 16 classes on the basis of the four parameters, viz. Associativeness, chronological appearance, frequency of occurrence and category. The taxonomy of keywords will enable to analyze research-trend of a subject and to identify potential research-areas of a subject.
10
Hajdu, Barat Agnes. "Multilevel education, training, traditions and research on UDC in Hungary." UDC Consortium, 2007. http://hdl.handle.net/10150/105607.
Full textAbstract:
This paper explores the theory and practice of education in schools and further education as two levels of the Information Society in Hungary. LIS education is considered the third level over previous ones. The curriculum and content of different subjects in schools and their relationship to libraries is summarized, as well as the training programmes for librarians, especially concerning knowledge organization. The long history of UDC usage in Hungary is surveyed, highlighting principal milestones and people. The paper provides a brief overview of recent developments, the situation after the new Hungarian edition, and current UDC usage and research directions.
11
Alabdullah, Salam Abdulqader Falih. "Classification of arcs in finite geometry and applications to operational research." Thesis, University of Sussex, 2018. http://sro.sussex.ac.uk/id/eprint/78268/.
Full textAbstract:
In PG(2; q), the projective plane over the field Fq of q elements, a (k; n)-arc is a set K of k points with at most n points on any line of the plane. When n = 2, a (k; 2)-arc is called a k-arc. A fundamental question is to determine the values of k for which K is complete, that is, not contained in a (k + 1; n)-arc. In particular, what is the largest value of k for a complete K, denoted by mn(2; q)? This thesis focusses on using some algorithms in Fortran and GAP to find large com- plete (k; n)-arcs in PG(2; q). A blocking set B is a set of points such that each line contains at least t points of B and some line contains exactly t points of B. Here, B is the complement of a (k; n)-arc K with t = q +1 - n. Non-existence of some (k; n)-arcs is proved for q = 19; 23; 43. Also, a new largest bound of complete (k; n)-arcs for prime q and n > (q-3)/2 is found. A new lower bound is proved for smallest size of complete (k; n)-arcs in PG(2; q). Five algorithms are explained and the classification of (k; n)- arcs is found for some values of n and q. High performance computing is an important part of this thesis, where Algorithm Five is used with OpenMP that reduces the time of implementation. Also, a (k; n)-arc K corresponds to a projective [k; n; d]q-code of length k, dimension n, and minimum distance d = k - n. Some applications of finite geometry to operational research are also explained.
12
Jeffreys, Christopher G. (Christopher Grey) 1979. "Support vector machine and parametric wavelet-based texture classification of stem cell images." Thesis, Massachusetts Institute of Technology, 2004. http://hdl.handle.net/1721.1/16651.
Full textAbstract:
Thesis (S.M.)--Massachusetts Institute of Technology, Sloan School of Management, Operations Research Center, 2004.Includes bibliographical references (p. 117-121).
This electronic version was submitted by the student author. The certified thesis is available in the Institute Archives and Special Collections.
Stem (cell research is one of the most promising and cutting-edge fields i the miedical sciences. It is believed that this innovative research will lead to life-saving treatments in the coming years. As part of their work, stem cell researchers must first determine which of their stem cell colonies are of sufficiently high quality to be suitable for experimental studies and therapeutic treatments. Since colony texture is a major discriminating feature in determining quality. we introduce a non-invasive, semi-automated texture-based stem cell colony classification methodology to aid researchers in colony quality control. We first consider the general problem of textural image segmentation. In a new approach to this problem. we characterize image texture by the subband energies of the image's wavelet decomposition, and we employ a non-parametric support vector machine to perform the classification that yields the segmentation. We also adapt a parametric wavelet-based classifier that utilizes the Kullback-Leibler distance. We apply both methods to a set of benchmark textural images, report low segmentation error rates and comment on the applicability of and tradeoffs between the non-parametric and parametric segmentation methods.
(cont.) We then apply the two classifiers to the segmentation of stem cell colony images into regions of varying quality. This provides stem cell researchers with a rich set of descriptive graphical representations of their colonies to aid in quality control. From these graphical representatiolns, we extract colony-wise textural features to which we add colony-wise border features. Taken together, these features characterize overall colony quality. Using these features as inputs to a multiclass support vector machine, we successfully categorize full stem cell colonies into several quality categories. This methodology provides stem cell researchers with a novel, non-invasive quantitative quality control tool.
by Christopher G. Jeffreys.
S.M.
13
San, Segundo Manuel Rosa. "Use of the UDC in Spain: implementation, application, teaching and research." UDC Consortium, 2007. http://hdl.handle.net/10150/106504.
Full textAbstract:
It was from 1895 onwards, the year in which the First International Bibliography Conference was held that the Decimal Classification system began to be implemented on a European scale and started to be disseminated in Spain. The introduction of the UDC scheme was initially subject to numerous difficulties faced by the various initiatives undertaken by individual librarians, but it subsequently received the support of the Spanish Administration. In 1939, the adoption of the UDC was established for all Spanish libraries, by official decree. Although what was stipulated in the decree was the 1934 German version, in practice libraries implemented the latest version of the UDC tables. Finally, from 1989 onwards, the compulsoriness of using UDC to classify collections and catalogues was abandoned, although its implementation in libraries, catalogues and bibliographies remained in most cases. The UDC is taught within the framework of regulated Library and Information Science schools, from both a theoretical and a practical point of view. In Spain, research on the UDC is already quite important: translations, adaptations and versions of the tables have been undertaken and there are also analytical works on different aspects of the UDC system.
14
Stalmans, Etienne Raymond. "DNS traffic based classifiers for the automatic classification of botnet domains." Thesis, Rhodes University, 2014. http://hdl.handle.net/10962/d1007739.
Full textAbstract:
Networks of maliciously compromised computers, known as botnets, consisting of thousands of hosts have emerged as a serious threat to Internet security in recent years. These compromised systems, under the control of an operator are used to steal data, distribute malware and spam, launch phishing attacks and in Distributed Denial-of-Service (DDoS) attacks. The operators of these botnets use Command and Control (C2) servers to communicate with the members of the botnet and send commands. The communications channels between the C2 nodes and endpoints have employed numerous detection avoidance mechanisms to prevent the shutdown of the C2 servers. Two prevalent detection avoidance techniques used by current botnets are algorithmically generated domain names and DNS Fast-Flux. The use of these mechanisms can however be observed and used to create distinct signatures that in turn can be used to detect DNS domains being used for C2 operation. This report details research conducted into the implementation of three classes of classification techniques that exploit these signatures in order to accurately detect botnet traffic. The techniques described make use of the traffic from DNS query responses created when members of a botnet try to contact the C2 servers. Traffic observation and categorisation is passive from the perspective of the communicating nodes. The first set of classifiers explored employ frequency analysis to detect the algorithmically generated domain names used by botnets. These were found to have a high degree of accuracy with a low false positive rate. The characteristics of Fast-Flux domains are used in the second set of classifiers. It is shown that using these characteristics Fast-Flux domains can be accurately identified and differentiated from legitimate domains (such as Content Distribution Networks exhibit similar behaviour). The final set of classifiers use spatial autocorrelation to detect Fast-Flux domains based on the geographic distribution of the botnet C2 servers to which the detected domains resolve. It is shown that botnet C2 servers can be detected solely based on their geographic location. This technique is shown to clearly distinguish between malicious and legitimate domains. The implemented classifiers are lightweight and use existing network traffic to detect botnets and thus do not require major architectural changes to the network. The performance impact of implementing classification of DNS traffic is examined and it is shown that the performance impact is at an acceptable level.
15
Goh, Siong Thye. "Machine learning approaches to challenging problems : interpretable imbalanced classification, interpretable density estimation, and causal inference." Thesis, Massachusetts Institute of Technology, 2018. http://hdl.handle.net/1721.1/119281.
Full textAbstract:
Thesis: Ph. D., Massachusetts Institute of Technology, Sloan School of Management, Operations Research Center, 2018.This electronic version was submitted by the student author. The certified thesis is available in the Institute Archives and Special Collections.
Cataloged from student-submitted PDF version of thesis.
Includes bibliographical references (pages 111-118).
In this thesis, I address three challenging machine-learning problems. The first problem that we address is the imbalanced data problem. We propose two algorithms to handle highly imbalanced classification problems. The first algorithm uses mixed integer programming to optimize a weighted balance between positive and negative class accuracies. The second method uses an approximation in order to assist with scalability. Specifically, it follows a characterize-then-discriminate approach. The positive class is first characterized by boxes, and then each box boundary becomes a separate discriminative classifier. This method is computationally advantageous because it can be easily parallelized, and considers only the relevant regions of the feature space. The second problem is a density estimation problem for categorical data sets. We present tree- and list- structured density estimation methods for binary/categorical data. We present three generative models, where the first one allows the user to specify the number of desired leaves in the tree within a Bayesian prior. The second model allows the user to specify the desired number of branches within the prior. The third model returns lists (rather than trees) and allows the user to specify the desired number of rules and the length of rules within the prior. Finally, we present a new machine learning approach to estimate personalized treatment effects in the classical potential outcomes framework with binary outcomes. Strictly, both treatment and control outcomes must be measured for each unit in order to perform supervised learning. However, in practice, only one outcome can be observed per unit. To overcome the problem that both treatment and control outcomes for the same unit are required for supervised learning, we propose surrogate loss functions that incorporate both treatment and control data. The new surrogates yield tighter bounds than the sum of the losses for the treatment and control groups. A specific choice of loss function, namely a type of hinge loss, yields a minimax support vector machine formulation. The resulting optimization problem requires the solution to only a single convex optimization problem, incorporating both treatment and control units, and it enables the kernel trick to be used to handle nonlinear (also non-parametric) estimation.
by Siong Thye Goh.
Ph. D.
16
Domm, Maryanne. "Mathematical programming in data mining: Models for binary classification with application to collusion detection in online gambling." Diss., The University of Arizona, 2003. http://hdl.handle.net/10150/280270.
Full textAbstract:
Data mining is a semi-automated technique to discover patterns and trends in large amounts of data and can be used to build statistical models to predict those patterns and trends. One type of prediction model is a classifier, which attempts to predict to which group a particular item belongs. An important binary classifier, the Support Vector Machine classifier, uses non-linear optimization to find a hyperplane separating the two classes of data. This classifier has been reformulated as a linear program and as a pure quadratic program. We propose two modeling extensions to the Support Vector Machine classifier. The first, the Linearized Proximal Support Vector Machine classifier, linearizes the objective function of the pure quadratic version. This reduces the importance the classifier places on outlying data points. The second extension improves the conceptual accuracy of the model. The Integer Support Vector Machine classifier uses binary indicator variables to indicate potential misclassification errors and minimizes these errors directly. Performance of both these new classifiers was evaluated on a simple two dimensional data set as well as on several data sets commonly used in the literature and was compared to the original classifiers. These classifiers were then used to build a model to detect collusion in online gambling. Collusion occurs when two or more players play differently against each other than against the rest of the players. Since their communication cannot be intercepted, collusion is easier for online gamblers. However, collusion can still be identified by examining the playing style of the colluding players. By analyzing the record of play from online poker, a model to predict whether a hand contains colluding players or not can be built. We found that these new classifiers performed about as well as previous classifiers and sometimes worse and sometimes better. We also found that one form of online collusion could be detected, but not perfectly.
17
Paknikar, Gayatri Suhas. "Facial Image Based Expression Classification System Using Committee Neural Networks." University of Akron / OhioLINK, 2008. http://rave.ohiolink.edu/etdc/view?acc_num=akron1210699575.
Full text
18
Meccio, Tony. "Feature Extraction and Randomized Learning for Image Analysis and Classification." Thesis, Università degli Studi di Catania, 2011. http://hdl.handle.net/10761/189.
Full textAbstract:
Research in Computer Vision has grown in interest over the years, up to the point where it became of interest for corporate research. The focus of this thesis is on studies which have been undertaken about designing Computer Vision applications for image analysis and classification in embedded systems. This research was funded by STMicroelectronics, AST Imaging, Catania Lab, and it is part of the activities of the Joint Lab between STMicroelectronics and the Image Processing Lab of the University of Catania.
19
Bellmore, Aimée Ryan. "Gender, Culture, and Prison Classification: Testing the Reliability and Validity of a Prison Classification System." PDXScholar, 2011. https://pdxscholar.library.pdx.edu/open_access_etds/423.
Full textAbstract:
Research consistently shows actuarial classification instruments have equal or higher predictive validity than clinical judgment and can lead to more ethical and fair treatment of incarcerated men and women (Austin, 1983, 1986; Bonta, 2002; Clements, 1981; Holsigner, Lowenkamp, & Latessa, 2006; Meehl, 1954; Salisbury, Van Voorhis, & Spiropoulos, 2009). Best correctional practice recommends all objective classification systems are tested for reliability and validity to ensure they are effective for the population they intend to serve (Austin, 1986; Holsinger et al., 2006; Salisbury et al., 2009). This study examined the reliability and validity of the classification and assessment instruments currently used by Golden Grove Adult Correctional Facility (Golden Grove), located on St. Croix in the United States Virgin Islands (USVI). Golden Grove is a mixed-gender, mixed-security status prison managed by the USVI territorial government, and is subject to United States Federal laws and mandates. Data from archival files were used to assess the internal reliability, construct validity, and predictive validity of the classification and assessment instruments used with incarcerated men and women at Golden Grove (N = 200). Primary objectives of this study were separated into four main categories: 1) examine the construct validity of Golden Grove's custody assessment tools; 2) investigate the predictive validity of Golden Grove's custody assessment tools across gender; 3) determine reliability and assess to what extent the primary classification officer's decisions have higher predictive validity than the actuarial tool; and 4) investigate the relationship between items on the needs assessment form and level of custody (minimum, medium, or maximum). Results were mixed but generally indicated weak reliability, construct validity, and predictive validity. Contrary to most research on gender and classification, a significant correlation between the initial custody score for incarcerated females and disciplinary reports (r = .26, n = 56, p < .05) indicated the initial custody tool predicted misconduct for maximum custody females better than for males. The mean number of disciplinary reports for maximum women (M = 1.12) was significantly higher compared to maximum men (M = .46). The classification officer overrode the instrument at a high rate for both the initial assessment instrument (44%) and the reassessment instrument (36.4%) rendering the objective assessment overly subjective. Overall, findings show the classification system at Golden Grove is not functioning as intended and improvements are recommended.
20
Wallace, Angus Keith, and wallace angus@gmail com. "Epilepsy research using nonlinear signal processing." Flinders University. Computer Science, Engineering and Mathematics, 2008. http://catalogue.flinders.edu.au./local/adt/public/adt-SFU20081124.210552.
Full textAbstract:
This thesis applies several standard nonlinear quantifiers to EEG analysis to examine both human primary generalised epilepsy (PGE) and rat models of human epilepsy.
We analysed rat EEG, and then used the analysed data, in parallel with an impedance recording, to better understand the events during experiments.
Next, the nonlinear analysis of EEG was used to attempt to model the behaviour of the impedance data. This modeling did not yield a useful predictive tool, so we recommend the continued recording of impedance data as a means of augmenting EEG recordings.
The analyses were also applied to human data, and showed differences between the PGE and control groups in apparently normal EEG. We then attempted to use these differences to detect the presence of PGE in an unclassified subject a diagnostic tool. This was done using a feed-forward neural network. We found that the inter-group differences were exploitable and facilitated the diagnosis of PGE in previously unknown subjects. The extent to which this is useful as a diagnostic tool should be assessed by further trials.
Finally, the analyses were used to examine data from a paralysed human subject, in an attempt to identify the mental task being performed by that
subject. This was not successful, suggesting that the same analyses that were useful in discriminating between PGE and control were not useful in detecting the mental state of the subject. It was also apparent that the presence of EMG (in an unparalysed state) assisted task-classification.
21
Kolli, Sam. "Classification of Research and Applications in Feature Modeling and Computer Aided Process Planning." Ohio University / OhioLINK, 2004. http://www.ohiolink.edu/etd/view.cgi?ohiou1108062913.
Full text
22
Blomkvist, Oscar. "Machine Learning Based Sentiment Classification of Text, with Application to Equity Research Reports." Thesis, KTH, Matematisk statistik, 2019. http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-257506.
Full textAbstract:
In this thesis, we analyse the sentiment in equity research reports written by analysts at Skandinaviska Enskilda Banken (SEB). We provide a description of established statistical and machine learning methods for classifying the sentiment in text documents as positive or negative. Specifically, a form of recurrent neural network known as long short-term memory (LSTM) is of interest. We investigate two different labelling regimes for generating training data from the reports. Benchmark classification accuracies are obtained using logistic regression models. Finally, two different word embedding models and bidirectional LSTMs of varying network size are implemented and compared to the benchmark results. We find that the logistic regression works well for one of the labelling approaches, and that the best LSTM models outperform it slightly.I denna rapport analyserar vi sentimentet, eller attityden, i aktieanalysrapporter skrivna av analytiker på Skandinaviska Enskilda Banken (SEB). Etablerade statistiska metoder och maskininlärningsmetoder för klassificering av sentimentet i textdokument som antingen positivt eller negativt presenteras. Vi är speciellt intresserade av en typ av rekurrent neuronnät känt som long short-term memory (LSTM). Vidare undersöker vi två olika scheman för att märka upp träningsdatan som genereras från rapporterna. Riktmärken för klassificeringsgraden erhålls med hjälp av logistisk regression. Slutligen implementeras två olika ordrepresentationsmodeller och dubbelriktad LSTM av varierande nätverksstorlek, och jämförs med riktmärkena. Vi finner att logistisk regression presterar bra för ett av märkningsschemana, och att LSTM har något bättre prestanda.
23
Gilbert, Mark, and Heather Lane. "Forty-five numbers for snow: a brief introduction to the UDC for Polar libraries." UDC Consortium The Hague, 2008. http://hdl.handle.net/10150/105332.
Full textAbstract:
This paper discusses the development of the Polar UDC. It examines some elements of the UDC specific to the Polar context, in particular the geographical auxiliary schedule. Some future plans for the implementation of UDC in a library and also in a museum context are outlined.
24
Rindelhardt, U., and F. P. Weiß. "Annual Report 2008 - Institute of Safety Research." Forschungszentrum Dresden-Rossendorf, 2009. https://hzdr.qucosa.de/id/qucosa%3A21602.
Full text
25
Krause, Wesley Allen. "An evaluation methodology using probation classification instruments in the selection of a nonequivalent control group." CSUSB ScholarWorks, 1989. https://scholarworks.lib.csusb.edu/etd-project/436.
Full text
26
Muketekelwa, Saliya L. "Investigating the potential of using hydrocyclone-fine screen hybrid systems to improve the performance of classification circuits." Master's thesis, University of Cape Town, 2017. http://hdl.handle.net/11427/27538.
Full textAbstract:
Classification is an integral part of comminution operations that controls the performance of the circuit. Hydrocyclones are normally used to perform the classification function. They offer numerous advantages that include, the ability to handle high throughputs, low floor space occupation and relatively low capital and running costs. Despite these advantages, hydrocyclones are inherently inefficient classifiers as they are predominantly dependent on hydrodynamics to effect separation. This effect is more prominent in operations handling complex ores such as a dual-density ore, where the heavy fine particles are misplaced to the underflow and the lighter middling particles report to the overflow. Several attempts have been made to improve the separation efficiency of cyclones either by modification of the cyclone or use of multi-stage cycloning. Most of the results obtained from experimental and simulation studies have shown considerable improvements. Even though some have not yet found wide application in the minerals industry due to practical limitations related to control and unstable operations. More recently, fine screening has gained recognition in the classification role. This development has allowed the use of fine screens in closed-circuit grinding operations resulting in significant metallurgical and economic benefits. Screens provide a sharper cut at the desired size and reduce the fraction of fines bypassing classification compared to hydrocyclones but have capacity limitations at smaller apertures. In an effort to mitigate the classification challenges of both the hydrocyclone and fine screen, this study investigated the potential of combining the high throughput performance of the hydrocyclone operation and the high precision classification characteristics of fine screening to result in a hybrid classification circuit Plant scale tests were conducted using five different classification circuit configurations at an operational Base Metal Concentrator treating a polymetallic ore. The classification circuit configurations considered included (i) a two-stage hydrocyclone with primary underflow reclassification (ii) an inclined hydrocyclone, (iii) a fine screen and (iv) selected permutations of hybrid circuit designs that included a hydrocyclone-fine screen (2 stage) and two hydrocyclones-fine screen (3 stage) variants of the hybridised configurations. The efficiency curves and their respective key performance indicators were used to assess the performance of the circuit configurations tested. The results showed that classification circuits that included fine screens exhibited higher sharpness of separation compared to circuit configurations comprised of hydrocyclones. The fine screen configuration showed the sharpest separation while the hydrocyclone-fine screen hybrid configurations gave relatively higher separation efficiencies than the configurations with hydrocyclones only. The overall sharpness of separation values obtained for the two stage and three-stage hybrid circuits were 3.0 and 2.4, respectively. The two-stage hydrocyclone and inclined hydrocyclone circuits had sharpness of separation values of 1.7 and 0.5, respectively. The inclined hydrocyclone circuit configuration performed the poorest. Furthermore, the two-stage hybrid circuit showed a higher degree of separation compared to the three-stage hybrid configuration. However, it was observed that a finer corrected cut size was realised for the three-stage hybrid circuit design. The fishhook effect was seen at particle sizes less than 38μm for the configurations incorporating a fine screen and an inclined hydrocyclone. Notably, the effect appeared to be more pronounced in configurations involving a fine screen stage. The results have shown that application of hybrid classification configurations can improve the performance of classification circuits. In addition, reclassification of hydrocyclone underflow on fine screens will results in a sharper classification while reclassifying the overflow stream on fine screens will provide a clean circuit final product. An evaluation of the capital and operating costs associated with fine screens should be done to determine the economic feasibility of incorporating the units in conventional milling circuits.
27
Sauerbrey, Roland, and Peter Joehnk. "discovered: THE HZDR RESEARCH MAGAZINE." Helmholtz-Zentrum Dresden-Rossendorf, 2012. https://hzdr.qucosa.de/id/qucosa%3A22183.
Full textAbstract:
\"Discovered\" is the English-language edition of our research magazine; it is published once a year. The magazine\'s German-language edition \"entdeckt\" is published biannually.
Each new issue of this easy-to-read magazine has a major focus, be it magnetic fields and forces, nuclear safety research, the DRESDEN-concept research alliance or cancer research. The magazine keeps you informed about research at the HZDR, new staff members or work groups as well as interesting events.
28
Rezvani, Zeinab, Johan Jansson, and Jan Bodin. "Advances in consumer electric vehicle adoption research: A review and research agenda." Elsevier, 2014. https://publish.fid-move.qucosa.de/id/qucosa%3A72831.
Full textAbstract:
In spite of the purported positive environmental consequences of electrifying the light duty vehicle fleet, the number of electric vehicles (EVs) in use is still insignificant. One reason for the modest adoption figures is that the mass acceptance of EVs to a large extent is reliant on consumers’ perception of EVs. This paper presents a comprehensive overview of the drivers for and barriers against consumer adoption of plug-in EVs, as well as an overview of the theoretical perspectives that have been utilized for understanding consumer intentions and adoption behavior towards EVs. In addition, we identify gaps and limitations in existing research and suggest areas in which future research would be able to contribute.
29
Ivanus, Lydia H. (Lydia Helen). "A case study of a patient classification system in a teaching hospital in Québec." Thesis, McGill University, 1995. http://digitool.Library.McGill.CA:80/R/?func=dbin-jump-full&object_id=23218.
Full textAbstract:
This study examines the application of the Project Research in Nursing (PRN) patient classification/workload measurement system in a teaching hospital in Quebec. It also compares the workload estimates of direct and indirect care activities determined from patients' total PRN point scores and from corresponding PRN levels of care (classes 1-15). This study examines the allocation and the rankings of patients' needs for nursing care in relation to the categories of the classification instrument. Two documentary sources, audit records of 1988-1989 to 1992-1993 and audit data collection forms of 1992-1993 were used to conduct the study and to provide a comparison of the results. Audit findings were relatively consistent over the five-year period. Comparison of pre-audit and post-audit workload estimates provided detailed information on the measurement sensitivity between total PRN point scores and PRN levels of care; the direction and magnitude of differences between classification ratings; and the distinctive nature of patients' needs for nursing care within and across clinical departments. This study discusses the implications of these findings for nursing resource management decisions, staffing and workload and makes recommendations for practice and further research.
30
Lagerström, Malin C. "Classification, Evolution, Pharmacology and Structure of G protein-coupled Receptors." Doctoral thesis, Uppsala University, Department of Neuroscience, 2006. http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-6356.
Full textAbstract:
G protein-coupled receptors (GPCR) are integral membrane proteins with seven α-helices that translate a remarkable diversity of signals into cellular responses. The superfamily of GPCRs is among the largest and most diverse protein families in vertebrates.
We have searched the human genome for GPCRs and show that the family includes approximately 800 proteins, which can divided into five main families; Glutamate, Rhodopsin, Adhesion, Frizzled/Taste2 and Secretin. This study represents one of the first overall road maps of the GPCR family in a mammalian genome. Moreover, we identified eight novel members of the human Adhesion family which are characterized by long N-termini with various domains. We also investigated the GPCR repertoire of the chicken genome, where we manually verified a total of 557 chicken GPCRs. We detected several specific expansions and deletions that may reflect some of the functional differences between human and chicken.
Substantial effort has been made over the years to find compounds that can bind and activate the melanocortin 4 receptor (MC4R). This receptor is involved in food intake and is thus an important target for antiobesity drugs. We used site-directed mutagenesis to insert micromolar affinity binding sites for zinc between transmembrane (TM) regions 2 and 3. We generated a molecular model of the human MC4R which suggests that a rotation of TM3 is important for activation of the MC4R.
Furthermore, we present seven new vertebrate prolactin releasing hormone receptors (PRLHRs) and show that they form two separate subtypes, PRLHR1 and PRLHR2. We performed a pharmacological characterization of the human PRLHR which showed that the receptor can bind neuropeptide Y (NPY) related ligands. We propose that an ancestral PRLH peptide has coevolved with a redundant NPY binding receptor, which then became PRLHR. This suggests how gene duplication events can lead to novel peptide ligand/receptor interactions and hence spur the evolution of new physiological functions.
31
Tóth, Valéria. "Hungarian digital toponym registry: results of a research programme." Deutsche Gesellschaft für Namenforschung (GfN), 2014. https://ul.qucosa.de/id/qucosa%3A13573.
Full textAbstract:
Das Digitale Ungarische Ortsverzeichnis. Ergebnisse eines Forschungsprojekts. – Im Sommer 2010 wurde unter dem Namen Digitales Ungarisches Ortsverzeichnis ein Forschungsvorhaben begonnen mit dem langfristigen Ziel der Aufnahme und Analyse des vollständigen Ortsnameninventars des Karpatenbeckens. Das Programm wurde in Zusammenarbeit mit Experten verschiedener ungarischer und ausländischer Einrichtungen aufgenommen, es möchte die Geschichte der Ortsnamen von den Anfängen bis heute untersuchen. Das Digitale Ungarische Ortsverzeichnis dient wissenschaftlichen Zwecken, kann aber gleichzeitig für ein allgemeines Publikum von Interesse sein. Die Datenbasis besteht aus zwei Teilen: das Moderne Ortsnamenverzeichnis enthält hauptsächlich Ortsnamen des 18. bis 20. Jahrhunderts, das Historische Ortsnamenverzeichnis erfasst Ortsnamen aus der Zeit bis 1350. Die zugrundeliegende Software-Architektur wird durch das 4D Datenbank-Management-System realisiert, das unter anderem über eine GIS- Komponente verfügt und damit die Visualisierung toponymischer Daten auf Karten ermöglicht: moderne toponymische Daten werden auf Fotografien von
Google Earth projiziert, während historische Daten auf rekonstruierten
mittelalterlichen Karten eingetragen werden. Dieser Beitrag möchte die allgemeinen und wissenschaftlichen Zielsetzungen des Digitalen Ungarischen Ortsverzeichnisses beschreiben und auf die bereits erreichten Resultate hinweisen. Das Ortsverzeichnis ist verfügbar
unter http://mdh.unideb.hu.
32
Tschirner, Erwin, and Olaf Bärenfänger. "Research Papers in Assessment." Institut für Testforschung und Testentwicklung e.V, 2015. https://ul.qucosa.de/id/qucosa%3A13924.
Full textAbstract:
Die "Research Papers in Assessment" sind eine Reihe des Instituts für Testforschung und Testentwicklung e.V. (ITT), in der Forschungsergebnisse, Tagungsbeiträge und wichtige Einzeldarstellungen veröffentlicht werden.
33
"Advances in Classification Research, Volume 17: Proceedings of the 17th ASIS&T SIG/CR Classification Research Workshop." dLIST, 2006. http://hdl.handle.net/10150/106360.
Full text
34
Verschelden, Mark Gerard. "Selecting research and development projects through classification." 1987. http://hdl.handle.net/2097/23777.
Full text
35
Kau, Lih-Jen, and 高立人. "A Research on Color Classification in Image System." Thesis, 1997. http://ndltd.ncl.edu.tw/handle/51927818844263402848.
Full textAbstract:
碩士國立交通大學
電機與控制工程學系
85
Human eye can classify different objects by way of their colors. The purpose of this research is to develop a color classification system in order to use in a self-made real time color image system. We hope to classify the target and background using their colors. The final system we made can process any object coated with multiple colors. Success of this research can also improve the defects or insufficiency of traditional gray image system. Fuzzy theory and back propagation neural network in conjunction with several typical color models to classify different colors were proposed. We proposed dynamic fuzzy input variable space partitioning in the fuzzy color classification system. We used single also double hidden layer neural network with adaptive learning rate in the back propagation neural network color classification system. A complete color classification system working under 32 bit operating system have been developed successfully. Parameters that will affect the performance of color classification were examined in this article. All experiments were finished under our self-made system. This research proposed image classification using colors. This is a new direction in color image processing. We still emphasized that this is just a beginning and new classification alogrithms should be found out to improve the performance or accuracy of color classification.
36
"Social relationship classification based on interaction data from smartphones." 2012. http://library.cuhk.edu.hk/record=b5549191.
Full textAbstract:
無線通信和移動技術已經從根本上改變了人和人之間相互通信的方式，隨著像智能手機這樣功能強大的移動設備不斷普及，現在我們有更多的機會去監測用戶的運動狀態、社交情況和地理位置等信息。近期，越來越多的基於智能手機的傳感研究相繼出現，這些研究利用智能手機中的多種傳感、定位以及近距離無線設備來識別手機用戶當前的活動狀態和周圍環境。一些可識別用戶活動狀態和監控身體健康狀況的移動應用程式已經被開發并投入使用。儘管如此，當前大部份關於智能手機的研究忽視了這樣一個問題，智能手機是用戶與外界通信的一個指令中心。移動用戶可以使用智能手機用很多種方式聯繫他們的朋友，例如打電話、發送短消息、電子郵件、或者通過即時通信程序或者社交網絡，這些多渠道的通信方式和人與人之間面對面的交流一樣重要，因此智能手機是識別用戶和其他聯繫人的社會關係的關鍵。在本論文中，我們提出用智能手機中 獨有的多渠道用戶通信數據來對用戶的的社會關係進行分類。作為我們研究的開始，我們生成人工的通信數據並且用社交矩陣來為人與人之間的通信建立模型，這也幫助我們測試了很多可以應用在此類問題的數據挖掘算法。接下來，我們通過招募真實用戶來採集他們的各種社交通信數據，這些數據包括手機通話記錄、電子郵件、社交網絡（Facebook和Renren）和面對面的交流。通過在社交矩陣上應用不同的分類算法，我們發現SVM的分類性能要超過KNN和決策樹算法，SVM對於社會關係的分類準確率可以達到82.4%。我們也對來自不同渠道的通信數據進行了比較，最終發現來自社交網絡和面對面交流的數據在社交關係分類中起更大的作用。另外，我們通過使用降低維度算法可以把社交矩陣從65維度映射到9維度，關係分類的準確率卻沒有明顯降低，在降低維度的過程中我們也可以提取出用戶主要的通信特徵，從而更好地解釋社會關係分類的原理。最後，我們也應用了CUR矩陣分解算法從社交矩陣65列中選出13列建立新的社交矩陣，關係分類的準確率從82.4%降低到77.7%，但是我們卻可以通過 CUR來選擇合適的傳感器抽樣採集頻率，這樣可以在利用手機採集數據過程中節省更多手機電量。Wireless Communications and Mobile Computing have fundamentally changed the way people interact and communicate with each other. The proliferation of powerful and programmable mobile devices, smartphones in particular, has offered an unprecedented opportunity to continuously monitor the physical, social and geographical activities of their users. Recently, much research has been done on smartphone-based sensing which leverages the rich set of sensing, positioning and short-range radio capabilities of the smartphones to identify the context of user activities and ambient environment conditions. Mobile applications for personal behavior tracking and physical wellness monitoring have also been developed. Despite that, most of the existing work in mobile sensing has neglected the role of smartphone as the command-center of the user’s communications with the outside world. As mobile users contact their friends via phone, SMS, emails, instant messaging, and other online social-networking applications, these multi-modal communication activities are as equally important as physical activities in proling one’s life. They also hold the key to understand the user’s social relationship with other people of interest. In this thesis, we propose to use the unique multi-model interaction data from smartphone to classify social relationships. To jump start our study, we generate articial interaction data and build social interaction matrix to modeMl the interaction between people. This also helps us in testing a wide range of data mining analysis techniques for this type of problem. We then carry out a social interaction data collection campaign with a group of real users to obtain real-life multi-modal communication data, e.g., phone call, Email, online social network(Facebook and Renren), and physical location/proximity. After applying different classification algorithms on social interaction matrix, we find that SVM outperforms KNN and decision tree algorithms, with a classification accuracy of 82.4% (the accuracies of KNN and decision tree are 79.9% and 77.6%, respectively). We also compare the data from different interaction channels and finally find that on-line social network and location/proximity data contribute more to the overall classification accuracy. Additionally, with dimensionality reduction algorithms, the social interaction matrix can be embedded from 65 to 9 dimensional space while preserving the high classification accuracy and we also get principle interaction features as by-product. At last, we use CUR decomposi¬tion to select 13 out of the 65 features in the social interaction matrix. The classification accuracy drops from 82.4% to 77.7% after CUR decomposition. But it can help to determine the right sensor sampling frequency so as to enhance energy efficiency for social data collection.
Detailed summary in vernacular field only.
Sun, Deyi.
Thesis (M.Phil.)--Chinese University of Hong Kong, 2012.
Includes bibliographical references (leaves 90-96).
Abstracts also in Chinese.
Chapter 1 --- Introduction --- p.1
Chapter 2 --- Research Background --- p.7
Chapter 2.1 --- Related work of social relationship analysis --- p.7
Chapter 2.1.1 --- Community detection in social network --- p.8
Chapter 2.1.2 --- Social influence analysis --- p.10
Chapter 2.1.3 --- Modeling social interaction data --- p.10
Chapter 2.1.4 --- Social relationship prediction --- p.12
Chapter 2.2 --- Classification methodologies --- p.14
Chapter 2.2.1 --- Algorithms for social relationship classification --- p.14
Chapter 2.2.2 --- Algorithms for dimensionality reduction --- p.16
Chapter 3 --- Problem Formulation of Relationship Classicification --- p.19
Chapter 3.1 --- Multi-modal data in smartphones --- p.20
Chapter 3.2 --- Formulation of relationship classification problem --- p.21
Chapter 3.3 --- Refinement of feature definition and energy efficiency --- p.27
Chapter 3.4 --- Chapter summary --- p.28
Chapter 4 --- Social Interaction Data Acquisition --- p.30
Chapter 4.1 --- Social interaction data collection campaign overview --- p.31
Chapter 4.2 --- Format of raw interaction data --- p.33
Chapter 4.3 --- Building social interaction matrix with real-life interaction data --- p.37
Chapter 4.4 --- Chapter summary --- p.43
Chapter 5 --- Statistical Analysis of Social Interaction Data --- p.45
Chapter 5.1 --- Coverage of social interaction data --- p.46
Chapter 5.2 --- Social relationships statistics --- p.48
Chapter 5.3 --- Social relationship interaction patterns --- p.52
Chapter 5.4 --- Chapter summary --- p.59
Chapter 6 --- Automatic Social Relationship Classification Based on Smartphone Interaction Data --- p.61
Chapter 6.1 --- Comparison of different classification algorithms --- p.62
Chapter 6.2 --- Advantages of multi-modal interaction data --- p.65
Chapter 6.3 --- Comparison of interaction data in different communication channels --- p.67
Chapter 6.4 --- Dimensionality reduction on social interaction data --- p.72
Chapter 6.5 --- Discussions in deploying social relationship classification application --- p.80
Chapter 6.5.1 --- Considerations of user privacy --- p.81
Chapter 6.5.2 --- Saving smartphone resources --- p.82
Chapter 6.6 --- Chapter summary --- p.83
Chapter 7 --- Conclusion and Future Work --- p.86
Bibliography --- p.90
37
Scribante, Juanett. "A patient classification system for critical care units." Thesis, 2014. http://hdl.handle.net/10210/9602.
Full textAbstract:
M.Cur.The nursing service manager is accountable for the managerial outputs of the nursing service, being quality nursing care, cost effectiveness and quality nursing staff management. These managerial responsibilities will influence the overall management efficiency of the health delivery system. There is a need for a more scientific basis of staffing in South African critical care units. Various methods are used to determine staffing needs, but the most accurate and acceptable method being used is the patient classification system. The success of a critical care patient classification system is based upon the accurate determination of the ideal nurse-patient ratio. Ideal nursepatient ratio depends on matching the competency of the critical care nurse to the severity of the critical care patient's illness. Internationally and nationally very few guidelines describe the competency requirements that the critical care nurse should have to function effectively in the critical care unit. To maintain a high standard of quality nursing care, the critical care nurse should be assigned to a critical care patient according to her competency level. The aim of this study was to describe a scientific patient classification system for critical care patients according to which workload and nursing requirements can be ascertained. Secondly, to describe guidelines for the competency requirements of a registered critical care nurse assigned to nurse the different categories of critical care patients categorised by the patient classification system as prescribed. The aim of the study is justified by means of an exploratory, descriptive, contextual research design. A critical care patient classification system has been developed and validated by a group of critical care nursing experts using a structured two phase research method. The statistical validity of each item of the critical care patient classification system was determined by means of a content validity index. All the items of the critical care patient classification system were rated as valid and therefore the critical care patient classification system as a whole has been accepted as valid. The reliability of the critical care patient classification system was tested by means of a pilot study in two selected research units. It was determined that the inter-rater reliability was within the 95 per cent confidence interval, that the system could discriminate between different critical care patients and nurse-patient ratios were effectively determined by the system. Critical care nurse competency requirements were identified by critical care nursing experts by means of a focus group interview. The results obtained from the focus group interview were verified by a literature control and peer group review. Guidelines for the competency requirements of the critical care nurse were compiled. The critical care patient classification system developed will equip the critical care manager with a scientific instrument to assign personnel to critical care units. The critical care patient classification system will furthermore assist the critical care unit manager in the daily utilisation of valuable human resources, in that critical care nurses with the appropriate level of expertise can be assigned to each patient, thus facilitating quality critical care nursing.
38
Chang, Gia-Hao, and 張家豪. "Research of Artificial Immune System Approach in MRI Classification." Thesis, 2007. http://ndltd.ncl.edu.tw/handle/07331574009535654866.
Full textAbstract:
碩士國立勤益科技大學
資訊與電能科技研究所
95
A lot of experts propose research about Artificial Immune System, and it has been widely used to a lot of fields, few applications are reported in medical image. This experiment is a brand-new application, and the purpose is to set up a kind of ais network suitable for medical image classification. Use the artificial immune system network to deal with the huge multi-spectral image data, meaning is the classification that the brain tissue. Finally, show the single tissue's image as result of the experiment. Offer reference information of tissue to a doctor quickly. More efficiently and more accurately. In order to verify the feasibility and efficiency of this method. We have used ROC (Receiver Operating Curve) method and Likelihood Ratios to assessment in many aspects. We compare with the Perceptron network of double layers, FCM method, prove AIS can be operated and fine.
39
SUNG, Hung－Kang, and 宋鴻康. "Research of ABC Classification Method upon Railway Inventory Management." Thesis, 2004. http://ndltd.ncl.edu.tw/handle/293579.
Full textAbstract:
碩士國立交通大學
管理學院碩士在職專班運輸物流組
92
This thesis focuses on the research of logistics and inventory management of railway warehousing, which are charged with supporting the maintenance of all types of rolling stock. The target of this integration is to reduce the quantity of parts held and to reduce the period material is held and reduce the time taken to pick components. This research relies on a case study to show some simple techniques developed in theory which can be applied to the railway’s current practice to improve the order-filling process, specifically consisting of ABC analysis as well as the relocation of the current inventory is simulated in the warehouse layout proposal. The layout of the warehouse with a total storage space occupied of roughly 16,159 square meters, with additional space for office use. Three main rooms have one aisle each with a width of 3 meters and 1,602 shelves of different sizes. The warehouse was focused with reducing the time needed to process orders especially when finding and picking the right components at the right time. To achieve a reduction in the process time we proposed the plan for improving the current practice, which aims to relocate all components, classified by the ABC analysis, so that these can then be retrieved more quickly and easily.
40
Hsu, Chieh-Yu, and 許婕渝. "Research of Using Classification Technique to Predict Overdue Credit." Thesis, 2011. http://ndltd.ncl.edu.tw/handle/21641220369165945554.
Full textAbstract:
碩士元智大學
資訊管理學系
99
As the year end or the annual tax season approaches, many banks increase advertising and promotion in personal loans for those in need of small loans. The revenue derived from these loans also became a main source of profit for these banks. This research exams the approved borrowers’ credit status from a sample bank. By using SPSS Clementine 12.0 data mining software to create sample models based on overall correct rate and overdue predictive rate, the best model will be determined based on classification model evaluation. The final chosen model will expedite lenders’ classification of small loan borrower and assist the underwriter in reduce lenders’credit risks.
41
Hwang, Yi Liang, and 黃益良. "Research of Neural Network Applied on Acoustic Signal Classification." Thesis, 1996. http://ndltd.ncl.edu.tw/handle/99087239492025619445.
Full textAbstract:
碩士中原大學
資訊工程研究所
84
For the widely adaptability of the Neural Networks in various applications, and the bright future of the development of appli- cation in Underwater Acoustic Signal. So, In this thesis, by combine the Neural Network and the Processing of the Underwater Acoustic Signal together, we try to discuss the ability of the Neural Network applied on the detection and recognition of the Passive Acoustic Signal, and the fitness for the Neural Network work in Underwater Acoustic Signal recognition. Use several different Neural Network Models-Back- Propagation Network, Fuzzy Adaptive Resonance Theory(Fuzzy ARTMAP) Neural Network and the Self-organization Map(SOM) Neural Network, after the processed of the Underwater Acoustic Signal, and use Fourier Transform to convert the signal to frequency domain then get the Feature Vector for this Signal. Use this Feature Vector as the input training data for the Neural Network. After the Neural Network trained, compare each Neural Network for the correct rate of the classification, then can decide which is the proper modal for Passive Sonar Signal recognition and classification. Finally, we promoting the rate of the classification for Passive Acoustic Signal, trying to accelerate the speed of the classification, reduce the learning times for the Neural Network by adjust the learning parameter or the activity function on the neuron unit and modify the architecture of the model of the Neural. Network we chosen in the research of this thesis.
42
Lin, Yu-hsin, and 林宇芯. "The research of organization classification with corporate social responsibility." Thesis, 2013. http://ndltd.ncl.edu.tw/handle/38287072748282508241.
Full textAbstract:
碩士國立臺南大學
科技管理碩士班
101
With the development of economy and society, for-profit organizations and enterprises are starting to focus on corporate social responsibility activities to contributing to our society, and fill the gap between the needs of society. In addition, corporate social responsibility is considered to enhance organizational competitiveness, and achieve sustainable development. This study investigated the degree of enterprise’s involvement on CSR-related activities, and which was divided into three dimensions, namely organizational, environment and social for analysis. Drawing on the “Corporate Social Responsibility”, “Organization classification”, and “Institutional theory” perspectives, this study aims to explore the differences and features of organization classification on the basis of corporate social responsibility .The finding of this research showed that enterprises under the three different kind of institutional pressure would operate CSR differently. Under a single pressure enterprises operate CSR-related activities with a lower willingness, whereas under multiple institutional pressures, including regulative pressure, normative pressure and cognitive pressure, the CSR-related activities operated by the studied companies have gradually improved.
43
ZHANG, Nai-Nan, and 張乃楠. "Research on Unbalanced Data Classification Based on Hybrid Method." Thesis, 2018. http://ndltd.ncl.edu.tw/handle/sy353q.
Full textAbstract:
碩士元智大學
資訊工程學系
107
Unbalanced data are ubiquitous in real-world datasets. In this paper, we investigate unbalanced data distribution for binary classification, i.e., where the number of majority class instances is significantly greater than the number of minority class instances. It is assumed that traditional machine learning algorithms attempt to minimize empirical risk factors, and, as a result, the classification accuracy of the minority is often sacrificed. However, people are often interested in the minority. Various data-level methods, such as over- and under-sampling, and algorithm-level methods, such as ensemble, cost-sensitive, and one-class learning, have been proposed to improve classifier performance with an unbalanced data distribution. Based on such methods, we proposed a hybrid approach to deal with unbalanced data problem that comprises data preprocessing, clustering, data balancing, model building, and ensemble.
44
Oosthuizen, Rika. "An evaluation of the PIE classification system in psychiatric social work." Thesis, 2012. http://hdl.handle.net/10210/6162.
Full textAbstract:
M.A.This study is born from the motivation to empower psychiatric social workers in South Africa with a classification system that compares well to classification systems used by other disciplines in the psychiatric setting. The classification system in question is the Person-in-Environment Classification System (PIE) and therefore this study sets out to investigate and report on the value the PIE might have for psychiatric social work in South Africa. Another objective then is to highlight what problems were experienced in using the PIE which will form the foundation for recommendations on further research and possible adaptations. The first part of this dissertation sets out the motivation and objectives of this study, where the second part gives a thorough description of psychiatric social work, the ecological approach in psychiatric social work, classification systems and the PIE itself. The research process is described whereafter the results of the data are explained. The results are discussed within the framework of three evaluative factors, namely image, applicability and structure which gives an indication of the value the PIE has for psychiatric social work. The conclusion can be drawn from this study that the PIE has definite value in all three evaluative factors. It can be concluded that the PIE can serve as an empowering tool in establishing and defining psychiatric social work's unique contribution to South African psychiatry.
45
Scheuer, Alice E. D. "A reconceptualization of paranoia : applications for research, classification, and intervention." Thesis, 1992. http://hdl.handle.net/10125/10220.
Full text
46
Shieh, Shih-Ping, and 謝仕屏. "Research and Implementation of Linux Netfilter extension for Packet Classification." Thesis, 2005. http://ndltd.ncl.edu.tw/handle/46545993081562475838.
Full textAbstract:
碩士逢甲大學
通訊工程所
93
Currently, many of the programs can take place using HTTP, TCP, and UDP over a dynamic port range; some even uses peer-to-peer (P2P) technology, creating problems in terms of network security. In my research, I will design and implement of the Linux Netfilter extension to classify packet. Through this extension module, we are able to analyze OSI Layer 7 packets, and are able to use the pre-define patterns to match the packet contents, hence, adding unique mark on the packet as it passes through the router. Finally, the marked packet will be recorded, discarded or sent to IMQ device for further process, such as bandwidth management. Netfilter will allow perfect combination of these two, hence creating better efficiency for packet comparison and categorization. This research has been tested on the SoftEther packet. Comparing the latency and loading information between the results of the ones that uses the extension module, and the ones that does not, it is proven the extension module can efficiently identify the SoftEther packets. And I will perhaps add more in-depth tests and other interesting scenarios: latency tests and real work load simultations.
47
Chen, Ching-Hsuan, and 陳璟萱. "The research about color of classification patterns in mobile application." Thesis, 2015. http://ndltd.ncl.edu.tw/handle/30076148356625813733.
Full textAbstract:
碩士國立交通大學
應用藝術研究所
103
For the past few years smartphones and tablets these mobile operating system are wildly used and shown off. More and more people rely on mobile operating system year after year. Under the situation, mobile operating system is not just for entertainment anymore but for daily assistant tools. The mobile application system can be the soul of the mobile operating system and recently is a big hit for designers to work on. Through App icon, users can interact with App and moreover the App icon elements consist with pattern and color. My research is mainly focus on color; I would like to find out if different App comes with its own color so I decide to propose this design idea. To discuss the question, first I should do an experiment through App based on pure pattern and pure color. And then I can realize the coordination between pattern and color is high or low. Final I use the sample which is combined two item (pattern and color) to do the suitable evaluation in order to reason out the influential to App classification with pattern and color. In the three experimental stages, I invited 30 Subjects to focus on pure pattern, pure color and the two combined icon questionnaire survey to do the evaluation. The result shows: (1) The statement is more concrete and the App icon pattern they used is more concrete as well. (2) Manipulating on App icon color, though now the trend of using orange and red colors is raising, however, the design entirely in blue or green colors can be easily judged as the fitter App category. (3) Photography category of colors has the formed tendency. (4) When imagine of icon pattern is very concrete, it’s hard for colors to effect the influence of icon imagine. On the contrary, when imagine of icon pattern is hard to associate with concrete idea, and colors will have a great influential effect. (5) In the same category some well-known App, the form and usage of colors have some extent influence in all.
48
Lin, Jin-Yi, and 林晋億. "Research of green coffee seeds classification using colour machine vision." Thesis, 2013. http://ndltd.ncl.edu.tw/handle/41821325846280181554.
Full textAbstract:
碩士國立雲林科技大學
電子與光電工程研究所碩士班
101
In order to classify the quality of coffee seeds, the thesis mainly discusses the system parameters collected then using color camera to recognize the color and the shape of coffee under the wavelength of white light. To obtain the parameters of samples we use image processing technology on image morphology. Finding out these parameters which have the biggest impact between these analytes. By means of combining selectors with back propagation artificial neural network, we construct a discrimination system to classify the qualiy of coffee seeds. Sketching the best color classification parameters system by using different chromaticity to coordinates such as (RGB), (HSV), (HSI), (HSL). Setting up a selector which can identify and collect color signals. Collecting parameters such as circumference, area and major axis from the center of mass.Minor axis from the center of mass, sleek rate, the number of selectors’ peak and bottom by using image morphology.Then using the method of learning back-propagation artificial neural network to classify the difference of shape. The part of color identification can achieve 100% success rate in experiment group. Artificial neural training network have 98.3% of identification success rate in experiment control group. Based on the analysis of experimental results, we could prove image processing mode which has a high accurate identification rate to classify the quality of coffee seeds. Therefore, the identification method of this system can set up parameters of screening and other functions automatic to distinguish the quality of coffee seeds.
49
Tsai, Ping-Yuan, and 蔡秉原. "Research of Behavior Classification Basedon Acoustic Features of Sperm Whale." Thesis, 2018. http://ndltd.ncl.edu.tw/handle/8w4q8k.
Full textAbstract:
碩士國立臺灣海洋大學
電機工程學系
106
Sperm whale possesses high intelligence and is the world's largest toothed cetacean mammal. It can communicate and establish social activities through different voices. By collecting sperm whales' sound data as well as performing feature analysis and classification, human beings can learn to understand the ecological behavior of sperm whales. In this thesis, we design an optimum classifier based on the maximum a posteriori probability (MAP) theory to classify the two types of sounds, Clicks and Codas, generated by sperm whales. We adopt these feature data for our research, which were constructed as a database through several years’ efforts and stored in the statistical signal processing laboratory, electrical engineering department of the national ocean university. Our proposed MAP classifier is used to analyze the classification accuracy of the sound characteristics after adding Gaussian white noise with different signal- to-noise ratios. The simulation results show that the classification accuracy can reach 100% for the signal-to-noise ratio being above a threshold of 20 dB.. Keywords: Sperm Whale, Acoustic Feature, maximum a posteriori probability (MAP)、Optimum Classifier
50
Wang, Decheng. "Stochastic modelling of magnetic resonance images with applications to tissue classification." 1996. https://scholarworks.umass.edu/dissertations/AAI9619452.
Full textAbstract:
This dissertation presents a new approach to stochastic modeling of magnetic resonance images. A rigorous and comprehensive model of MR image formation is derived based on the physics of magnetic resonance image formation, and the model is validated using real MRI data. A general theoretical result about the existence of spatial autoregressive processes is presented. We attempted to obtain an accurate tissue classification map for a set of real MR images based on two doctors' hand tracings of the boundaries of the regions of the different tissues in the images. In accordance with the prior knowledge of the tissue map, we designed a prior model for the distribution of the underlying tissue map. The technique is applied to tissue classification on a set of real MR images by the use of the Bayesian formulation and ICM algorithm.