Dissertations / Theses on the topic 'Source profiling'

Thesis: M.B.A., Massachusetts Institute of Technology, Sloan School of Management, 2015. In conjunction with the Leaders for Global Operations Program at MIT.
Thesis: S.M., Massachusetts Institute of Technology, Engineering Systems Division, 2015. In conjunction with the Leaders for Global Operations Program at MIT.
Title as it appears in MIT Commencement Exercises program, June 5, 2015: Source base optimization : footwear profiling at Nike, Inc. Cataloged from PDF version of thesis.
Includes bibliographical references (page 31).
The long term sourcing for footwear development, or "profiling," at Nike, Inc. has grown with the company and become significantly complex. It is no longer possible for a single person, no matter the level of experience, to optimize the company's profiling plan without computational assistance. Optimization methods, specifically mixed-integer linear programing, present an opportunity to save between 6.7 and 9.7% of combined labor and duty costs to the company. The model proposed by this research is responsible for justifying that potential but is merely a starting point for Nike, Inc. Further application and research into the company's manufacturing processes including transportation costs, technology groupings, and the Manufacturing Index (MI) could wield results that far surpass the levels obtained by this research. Implementation of an algorithmic approach is challenging for an organization that values "storytelling," collaboration, and narrative. However, in time I believe that this model, or something similar, will find a place, and deliver results, for Nike, Inc.
by David G. Jacobs.
M.B.A.
S.M.

The market needs for embedded or mobile devices were exploding in the last few years. Customers demand for devices that not only have high capacity of managing various complex jobs, but also can do it fast. Manufacturers therefore, are looking for a new field of processors that fits the special needs of embedded market, for example low power consumption, highly integrated with most components, but also provides the ability to handle different use cases. The traditional ASICs satisfied the market with great performance-per-watt but limited scalability. ASIP processors on the other hand, impact the new market with the ability of high-speed optimized general computing while energy efficiency is only slightly lower than ASICs. One essential problem in ASIP design is how to find the algorithms that can be accelerated. Hardware engineers used to optimize the instruction set manually. But with the toolset introduced in this thesis, design automation can be made by program profiling and the development cycle can be trimmed therefore reducing the cost. Profiling is the process of exposing critical parts of a certain program via static code analysis or dynamic performance analysis. This thesis introduced a code profiler that targeted at discovering repetition section of a program through static and dynamic analysis. The profiler also measures the payload of each loop and provides profiling report with a user friendly GUI client.

Technological advancements in mass spectrometry allowing quantification of almost complete proteomes make proteomics a key platform for generating unique functional molecular data. Furthermore, the integrative analysis of genomic and proteomic data, termed proteogenomics, has emerged as a new field revealing insights into gene expression regulation, cell signalling, and disease processes. However, the lack of software tools for high-throughput integration and unbiased modification and variant detection hinder efforts for large-scale proteogenomics studies. The main objectives of this work are to address these issues by developing and applying new software tools and data analysis methods. Firstly, I address mapping of peptide sequences to reference genomes. I introduce a novel tool for high-throughput mapping and highlight its unique features facilitating quantitative and post-translational modification mapping alongside accounting for amino acid substitutions. The performance is benchmarked. Furthermore, I offer an additional tool that permits generation of web accessible hubs of genome wide mappings. To enable unbiased identification of post-translational modifications and amino acid substitutions for high resolution mass spectrometry data, I present algorithmic updates the mass tolerant blind spectrum comparison tool ’MS SMiV’. I demonstrate the applicability of the changes by benchmarking against a published mass tolerant database search of a high resolution tandem mass spectrometry dataset. I then present the application of ‘MS SMiV’ on a panel of 50 colorectal cancer cell lines. I show that the adaption of ‘MS SMiV’ outperforms traditional sequence database based identification of single amino acid variants. Furthermore, I highlight the utility of mass tolerant spectrum matching in combination with isobaric labelled quantitative proteomics in distinguishing between post-translational modifications and amino acid variants of similar mass. In the last part of this work I integrate both tools with a high-throughput proteogenomic identification pipeline and apply it to a pilot study of chondrocytes derived from 12 osteoarthritic individuals. I show the value of this approach in identifying variation between individuals and molecular levels and highlight them with individual examples. I show that multi-plexed proteogenomics can be used to infer genotypes of individuals.

In forensic casework analysis it is necessary to obtain genetic profiles from increasingly smaller amounts of biological material left behind by perpetrators of crime. The ability to obtain profiles from trace biological evidence is demonstrated with so-called 'touch DNA evidence' which is perceived to be the result of DNA obtained from shed skin cells transferred from donor to an object or person during physical contact. However, the current method of recovery of trace DNA involves cotton swabs or adhesive tape to sample an area of interest. This "blind-swabbing" approach may result in the recovery of biological material from different individuals resulting in admixed DNA profiles which are often difficult to interpret. Profiles recovered from these samples are reported to be from shed skin cells with no biological basis for that determination. A specialized approach for the isolation of single or few cells from 'touch DNA evidence' is necessary to improve the analysis and interpretation of recovered profiles. Here we describe the development of optimized and robust micro volume PCR reactions (1-5 uL) to improve the sensitivity and efficiency of 'touch DNA' analysis. These methods will permit not only the recovery of the genetic profile of the donor of the biological material, but permit an identification of the tissue source of origin using mRNA profiling. Results showed that the 3.5 uL amplification volume, a fraction of the standard 25 uL amplification volume, was the most ideal volume for the DNA assay, as it had very minimal evaporation with a 50% profile recovery rate at a single cell equivalent input (~5 pg) with reducing amplification volume alone. Findings for RNA showed that by reducing both amplification steps, reverse transcriptase PCR (20 uL) and body fluid multiplex PCR (25 uL), to 5 uL, ideal results were obtained with an increase in sensitivity and detection of six different body fluids down to 50 pg. Once optimized at the trace level, the assays were applied to the collection of single and few cells. DNA findings showed that about 40% of a full profile could be recovered from a single buccal cell, with nearly 80% of a full profile recovered from only two cells. RNA findings from collected skin particles of "touched" surfaces showed accurate skin detection down to 25 particles and detection in one clump of particles. The profiles recovered were of high quality and similar results were able to be replicated through subsequent experiments. More studies are currently underway to optimize these developed assays to increase profile recovery at the single cell level. Methods of doing so include comparing different locations on touched surfaces for highest bio-particle recovery and the development of physical characteristics of bio-particles that would provide the most ideal results.
M.S.
Masters
Chemistry
Sciences
Forensic Science; Forensic Biochemistry Track

L’évolution de l’Internet dans les dernières années a été caractérisée par des changement dramatiques dans la manière dont les utilisateurs se comportent, interagissent et utilisent le réseau. Ceci a été particulièrement accompagné par l’introduction de nouvelles classes d’applications telles que les jeux en ligne et les réseaux pair-à-pair. L’un des défis les plus importants pour les administrateurs réseau et les ISPs est alors devenu l’identification du trafic Internet afin de pouvoir protéger leurs ressources contre le trafic indésirable et de prioriser certaines applications majeures. Les méthodes statistiques sont préférées à celles basées sur le numéro de port et l’inspection approfondie des paquets, car elles sont robustes au changement malveillant du numéro de port et fonctionnent avec le trafic crypté. Ces méthodes combinent l’analyse des paramètres statistiques des flux de paquets, tels que la taille des paquets et le temps les séparant, avec des techniques issues de la théorie d’apprentissage (machine learning). La majorité des méthodes statistiques ne peuvent pas identifier les flux applicatifs en temps réel et elles ont besoin d’atteindre la fin des flux avant de prendre une décision sur leur nature. Ceci est considéré comme trop long pour la plupart des administrateurs réseau, puisqu’il ne permet pas de bloquer un flux Internet indésirable à son début ni de lui donner en amont une qualité particulière de service. Un autre défi important pour les administrateurs réseau est de détecter et diagnostiquer tout changement dans le réseau comme une congestion à long terme, un changement dans le routage, une défaillance d’une liaison ou tout autre événement entraînant un changement dans les délais réseau. Dans la littérature, il y a un grand nombre de méthodes pour détecter des anomalies dans le réseau, mais la plupart de ces méthodes ont besoin de générer un volume considérable de trafic destiné) la métrologie du réseau. La réduction de la charge des mesures est un besoin vital pour les administrateurs réseaux. Dans cette thèse, nous décrivons les travaux que nous avons menés sur l’identification du trafic Internet et sur la détection des anomalies dans les réseaux. Dans la première partie, nous présentons nos trois méthodes que nous avons développées au cours de cette thèse, et qui permettent d’identifier avec précision et à la volée le trafic Internet. La première méthode, par sa nature itérative et probabiliste, identifie les applications rapidement et avec une grande précision en utilisant uniquement la taille des N premiers paquets. La deuxième méthode enrichit la première avec les temps entre paquets, pour cela nous avons eu besoin d’introduire un modèle pour filtrer le bruit dû aux conditions du réseau et d’extraire des mesures le temps d’attente due aux applications. Notre troisième méthode pour la classification du trafic en ligne combine les approches statistiques à des informations sur le comportement des machines hôtes afin de rendre l’identification du trafic Internet encore plus précis tout en profilant les activités réseaux des hôtes. Pour notre troisième méthode, nous utilisons la taille des paquets comme paramètre principal et nous exploitons les informations sur l’interaction des machines pour mieux affecter un flux à une application. Dans la deuxième partie de cette thèse, nous abordons le problème de détection des anomalies dans les réseaux. Nous commençons par une étude sur la stabilité des systèmes de coordonnées Internet (particulier Vivaldi). Dans une première étape, nous confirmons le fait que les coordonnées de Vivaldi oscillent au fil du temps en raison de la nature adaptative du système. Toutefois, les variations de ces coordonnées sont dans la plupart des temps en corrélation les unes avec les autres, pointant par conséquent vers un cluster de nœuds stables vu de l’intérieur du réseau. Dans un deuxième temps, nous présentons un nouvel algorithme de cloustering basé sur des méthodes de groupement hiérarchique afin d’identifier ce cluster de nœuds stables. Enfin, nous soulignons l’utilité d’une telle constatation avec une application qui permet de détecter les changements dans le réseau. En changeant artificiellement les délais du réseau dans différents scénarios, nous montrons que ces changements sont reflétés par ce corps de nœuds stables, permettant ainsi d’obtenir une image globale de la stabilité du réseau sans avoir besoin de mesures exhaustives des délais
The evolution of the Internet in the last few year has been characterized by dramatic changes in the way users behave, interact and utilize the network. This was accompanied by the introduction of new categories of applications such as network games and peer-to-peer services. One of the most important challenges for network administrators and ISPs is then becoming the identification of Internet traffic applications in order to protect their resources from unwanted traffic and to prioritize some major applications. Statistical methods are preferred to port-based ones and deep packet inspection since they donât rely on the port number and they also work for encrypted traffic. Theses methods combine the statistical analysis of the application packet flow parameters, such as packet size and inter-packet time, with machine learning techniques. However, the majority of these statistical methods cannot identify flows early and require reaching the end of flows before taking any decision which is considered as too late for network administrators ; indeed they do not provide means to stop an Internet flow or to give it a special quality of service early in its lifetime. Another important challenge for network administrators is to detect and diagnose key network changes as a long-term congestion, a rerouting, a link failure or any other even causing a shift in network delays. In the literature there is a huge amount of anomaly detection methods but most of them require exhaustive measurements to function properly. Reducing the load of network-wide monitoring is always a vital need for network administrators. In this thesis we present several contributions around Internet traffic identification and network-wide anomaly detection. In the first part we present three methods we have developed in order to identify accurately and on the fly the Internet traffic. The first method is a new online iterative probabilistic method that identifies applications quickly and accurately by only using the size of the first N packets. The second method enhances the first one with the inter-packet time in order to identify Internet traffic, this has required the introduction of a model to isolate the noise due to network conditions and to extract the time generated by the applications themselves. Our third method is a new online method for traffic classification that combines the statistical and host-based approaches in order to construct a robust and precise method for early Internet traffic identification. We use the packet size as the main feature for the classification and we benefit from the traffic profile of the host (i. E. Which application and how much) to decide in favour of this or that application. In the second part of this thesis, we aboard the problem of network-wide anomaly detection. We start by making a study about the stability of Internet coordinate systems (especially Vivaldi). In a first stage we confirm the fact that Vivaldi coordinates are most of the time correlated with each other pointing to a stable cluster of nodes seen from inside the network. In a second stage, we present a new clustering algorithm based on the data mining Hierarchical Grouping method to identify this cluster of stable nodes. Finally, we highlight the utility of such finding with and application that tracks changes in network delays. By changing artificially the network delays in different scenarios, we show that these changes are easily reflected by this body of stable nodes, hence allowing to obtain a global picture about the stability of the underlying network without the need for exhaustive delay measurements

The presence of faecal material in rivers and bathing waters can degrade water quality and impact on drinking water supply, fishing, aquaculture and recreational uses. In developing countries, infectious diseases related to the faecal contamination of water resources can constitute a major burden on human health. Determining the origin of contamination is not only imp0l1ant when assessing the degree of risk posed to public health, but also in deciding how best to target remediation measures. This study set out to develop and implement a low-cost antibiotic resistance analysis (ARA) method for determining the sources of environments. 2 ~ OC1 2006 aquatic SERV\CES ARA is a potentially useful method for distinguishing faecal ba~ri~ttOG~U1J'Cre r~1M has been used successfully to predict the origin of faecal contaminaQ n in rIver t tchments in the U.S.A. This phenotypic approach has cost benefits over genotypic approaches but existing protocols are time-consuming and manual data handling is open to human error. A simplified, low-cost approach to the ARA technique was developed that uses automated data recording combined with simple statistical analyses to compare bacteria of the genus Enterococcus from various faecal sources. In total, 3675 isolates were analysed for their sensitivity to a combination of antibiotic concentrations, creating a database of over 250,000 test reactions. Discriminant function analysis (DF A) revealed how many of the isolates were correctly grouped into each source category, on the basis of their antibiotic resistance patterns. When the database was split into ' municipal wastewater (MW) vs. livestock vs. wild bird' 73% the database isolates were correctly classified. This rate fell slightly to 68% when the database was split into 'MW vs. cattle and sheep vs. pigs vs. wild birds vs. dogs vs. poultry'. Isolates of porcine origin demonstrated the highest levels of resistance to the chosen antibiotic concentrations and the best rates of classification. Once the most discriminant antibiotics had been identified, it was possible to reduce the number of concentrations from 80 to just 31 . The most common sources of faecal pollution present in river water samples were livestock and wild birds. Whilst isolates classified as originating from municipal wastewaters were also detected, they were not found to be the dominant source. The data presented in this thesis also suggest that patterns of antibiotic resistance amongst bacterial populations differ with respect to geographical location. Future faecal source tracking studies using ARA may therefore require the assembly of catchment specific databases. However, antibiotic resistance patterns did appear to remain stable for at least 36 months. Most imp0l1antly, the ARA technique has demonstrated that it is capable of distinguishing the human and animal populations responsible for a faecal pollution incident, thereby allowing a more cost-effective and targeted approach to the remediation of impaired surface waters in the future.

Mon projet a eu pour ambition de progresser dans la compréhension des mécanismes cellulaires et moléculaires qui sous-tendent les multiples effets médiés par l’oestradiol et les récepteurs des oestrogènes. La dissection des mécanismes de signalisation a passé par l’étude de l’expression génique à grande échelle (microarray) dans certains modèles de souris que nous avons développés. Nous disposions plus particulièrement de modèles de souris:1- porteuses de mutations germinales pour ERa et ERb (ERKOs); 2- porteuses de la mutation germinale pour l’aromatase (ArKO). Notre projet devait nous permettre i) - d’identifier des gènes dont l’expression est controlée par l’oestradiol et/ou par les récepteurs nucléaires des oestrogènes ii) - d’évaluer les effets oestrogène-dépendants mais oestrogènes récepteurs indépendants (récepteurs membranaires) d’une part et de les distinguer des évènements oestrogène-indépendants mais oestrogènes récepteurs dépendants d’autre part. Les étapes de l’approche sont résumées comme suit : valorisation des résultats [détection du signal, analyse des données brutes (expression différentielle, normalisation, filtrage, statistiques, regroupement)], vérification et confirmation des données issues de la puce à ADN par PCR, interprétation biologique des données. En conclusion : les informations obtenues démontrent que les résultats des expériences de microarrays permettent de sélectionner des gènes de manière reproductible et non-ambigüe. Dans les tissus étudiés (prostate et tissu adipeux blanc) les analyses du profil d’expression de ces gènes révèlent que les effets sont principalement médiés par le récepteur alpha des oetrogènes et de manière indépendante de l’hormone
The definition of the role of estrogen is a long-date scientific preoccupation. On the basis of the studies carried out in the last twenty years, it is now well accepted that estradiol and its cognate receptors are relevant transcription regulators in reproductive as well as non-reproductive tissues. Models of estrogen (E2) insufficiency (ArKO) and estrogen receptors (ER) dysfunction (ERKOs) have revealed new and unexpected roles of estradiol and its receptors both in female and male. The purpose of this study is to use mouse models of estrogen insufficiency (ArKO) and estrogen receptors dysfunction (ERKOs) to provide a genomic insight in the multiple and complex mechanisms defining estrogenic signaling to help understanding its role in physiological and pathological conditions. In particular the objective was to identify the genes responsive to estrogen signaling according to various possible mechanisms: 1) estrogen and estrogen receptordependent actions; 2) estrogen-independent and estrogen-receptor-dependent actions; 3) estrogen receptor-independent estrogen-dependent actions. To reach this aim, estrogen and estrogen receptors dependent genes expression profiling were performed by microarray analysis in ventral and dorso-lateral prostate and gonadal white adipose tissue from mouse models of impaired estrogen synthesis (ArKO) and ER action (ERKOs). The experimental and biological reproducibility of microarray data was first verified and confirmed providing a correlation between real-time PCR and microarrays in fold change measurements and in expression profiles across all tissues. The results obtained from the analysis of the expression profiles indicate that the classical and the non-genomic actions of estrogen are not to represent the main mechanisms of estrogenic signaling in prostate and adipose tissue. Conversely it appears that the estrogenic signaling in these tissues is exerted via estrogen receptors with an estrogen-independent mechanism of action. ERa appear to be the main mediator of the observed estrogenic effects, with mechanisms that differ according to the specific tissue. In ventral and dorso-lateral prostates, ERa seems to have inhibitory effects on transcription of target genes, supporting the hypothesis of its implication as a tumor suppressor in the prostate gland. Additional studies need to be performed to permit the identification of genes whose regulation can be directly modulated by ERs

Les mutations hétérozygotes du gène codant pour le facteur de transcription HNF1B sont à l'origine d'un syndrome multisystémique complexe connu sous le nom de « Renal Cysts and Diabetes » (RCAD). Un modèle de souris généré dans notre laboratoire s'est avéré reproduire plusieurs caractéristiques de la maladie humaine. Nous avons réalisé un séquençage ARNm-microARN à différents stades de développement (E14,5 ; E15,5 ; E17,5) de ce modèle. Nous avons montré que les gènes les plus dérégulés étaient impliqués dans les processus métaboliques de transport, de lipides et d’acides organiques et étaient exprimés dans les tubules proximaux et, dans une moindre mesure, dans l’anse de Henlé et les canaux collecteurs. Nous avons sélectionné quatre microARN (miR-802, 194-2, 192 et -30a), régulés à la baisse et potentiellement contrôlés par HNF1B. Des expériences de transactivation de gène rapporteur dans des cellules HEK-293 ont montré que HNF1B était capable de transactiver la transcription de ces microARN via des sites de liaison présents dans les séquences régulatrices de ces gènes. En utilisant des microARN MIMICS nous avons par la suite montré que mir-802, mir-194-2 et mir-192 étaient capables d'inhiber l’expression d’un gène rapporteur contenant la région 3'UTR de HNF1B. L'analyse d'échantillons d'urine de 22 patients RCAD et de 22 contrôles sains a permis d'identifier 146 peptides excrétés de manière différentielle et associés au syndrome. En utilisant ces résultats dans un modèle mathématique, classificateur prédit efficacement le syndrome RCAD avec une sensibilité de 91.7% et une spécificité de 91.1% sur une large population de patients
Heterozygous mutations in the gene encoding the transcription factor HNF1B are the cause of a complex multisystem syndrome known as Renal Cysts And Diabetes (RCAD). A mouse model generated in our laboratory was shown to reproduce several features of the human disease. We performed high-throughput mRNA-microRNA sequencing at different developmental stages (E14.5, E15.5, E17.5). We showed that the most down-regulated genes were involved in transport, lipid and organic acid metabolic processes and expressed in proximal tubules and to a lesser extent in the loop of Henle and collecting ducts. We then selected four microRNAs (mir-802, 194-2, 192 and -30a), which were down-regulated and potentially controlled by HNF1B. Luciferase assays in HEK-293 cells showed that HNF1B was able to specifically transactivate in a dose response mode these microRNAs through binding HNF1B-binding sites in their regulatory promoter/enhancer upstream sequences. We subsequently showed by luciferase assays using miRNA MIMICS that mir-802, mir-194-2 and mir-192 were able to inhibit luciferase vectors containing the 3’UTR of Hnf1b. Analysis of urine samples from 22 RCAD patients and 22 healthy controls led to the identification of 146 peptides differentially excreted and associated with RCAD including a similarity regarding collagen and uromodulin fragments with the RCAD mouse model. Combining the peptides into a mathematical model we used independent cohorts of patients to validate the prediction of the RCAD syndrome. Our classifier efficiently predicted RCAD syndrome with 91.7% sensitivity and 91.1% specificity on a wide population

 U okviru ove doktorske disertacije posmatrani su zakoni održanja sa funkcijom fluksa koja ima prekid u x = 0, pri čemu delovi fluksa levo i desno od x = 0 imaju smo po jedan ekstrem. U prvoj glavi se može naći pregled osnovnih pojmova, definicija i teorema. U drugoj  glavi su opisani hiperbolični sistemi zakona održanja, slaba rešenja, kao i numerički postupci za njihovo rešavanje. U trećoj glavi su predstavljeni  diskretni profili darnih talasa. U četvrtoj glavi su opisani zakoni održanja  sa prekidnom funkcijom fluksa i ukratko su predstvaljeni rezultati drugih autora iz ove oblasti. U petoj glavi je najpre analizirana tzv. jednačina sa dva fluksa u slučaju kada oba dela fluksa levo i desno od  x = 0 imaju minimum, a pri tome se seku u najviše jednoj tačci unutar intervala. Primenom regularizacije na intervalu [−ε, ε], za ε > 0 dovoljno malo, dokazano je postojanje diskretnih udarnih profila za postupak Godunova za zakone održanja sa promenljivom funkcijom fluksa. Definisan je i odgovarajući diskretan uslov entropije, a postojanje entropijskog diskretnog  udarnog profila je postavljen kao kriterijum za dopustivost udarnih talasa. Potom je analizirana ista jednačina u slucaju kada deo fluksa levo  od x = 0 ima maksimum, a deo fluksa desno od x = 0 minimum, dok se oba dela fluksa seku na  krajevima posmatranog intervala. U ovom slučaju, uopšten je uslov entropije. U okviru ove glave je prikazano nekoliko numeričkih primera za oba opisana slučaja. Numerički rezultati  su dobijeni korišcenjem softvera razvijenog za potrebe ove teze u programskom paketu Mathematica.
We consider conservation laws with a flux discontinuity at x = 0, where the flux parts from both left and right hand side of x = 0 have at most one extreme on the  observed  domain. The first chapter provides elementary definitions and theorems..The second chapter refers to hyperbolic systems of conservation laws, their solutions, and  numerical procedures. The third chapter is devoted to discrete  shock profiles. The fourth chapter describes conservation laws with discontinuous flux functions and provides basic information upon known results in this field. In the  fifth chapter, we first  analyse the two-flux equation when both flux parts have a minimum and cross at most    at one point in the interior of the domain. Using a flux regularization on the interval [−ε,   ε], for ε > 0 small enough, we show the existence of discrete shock profiles for Godunov’s scheme for conservation laws with discontinuous flux functions. We also define a discrete entropy condition accordingly, and use the existence of an entropy discrete shock profile as an entropy criterion for shocks. Then we analyse the same problem in the case when the flux part on the left of x = 0 has a maximum and the part on the right of x = 0 has a minimum, whereas the fluxes cross at the edges of the interval. We derive a more general discrete entropy condition in this case. We provide several numerical examples in both of the above mentioned flux cases. All the  presented numerical results are obtained using a program written in Mathematica. Finally, in chapter six, we prove the existence of  singular shock waves in the case when the graph of one of the flux parts is above the graph of the other one on the entire domain. For that purpose, we use the shadow wave technique. At the end of this chapter, we provide a numerical verification of the obtained singular solution.

Ерозија земљишта је један од најкомплекснијих проблема и уједно један од најподмуклијих човекових непријатеља. Ерозивни процеси постепено разарају и уништавају земљиште, односе хранљиве супстрате и ремете водни режим. Они доводе становништво многих, често пространих области у беду, сиромаштво и миграцију. Због тога, многе привредне гране морају бити увучене у борбу против ерозије земљишта и њених страховитих последица. Ерозија је веома сложена динамичка појава, која зависи од низа фактора чије је диференцирање врло тешко. Испитивање и утврђивање интензитета водне ерозије спада у групу проблема који се не могу свеобухватно решити. Веома је тешко, скоро немогуће, издвојити све факторе, а потом мултиплицирати све те факторе у свим постојећим комбинацијама. Овим радом сагледано је стање ерозије земљишта у сливу Млаве и Пека, односно Хомољу и Звижду, са више аспеката и утврђен низ појава као што су: концентрација растворених минералних материја, процентуално учешће макроелемената у укупној минерализацији, однос минерализације према протицају, укупан годишњи пронос растворених минералних материја, однос специфичне електричне проводљивости и укупне минерализације, износ хемијске евакуације, износ хемијски растворених минералних материја које потичу од стенског комлекса у сливу, утицај геолошког састава на укупну минерализацију воде, њену структуру и специфичну хемијску ерозију, концентрација суспендованог наноса по профилима, однос суспендованог наноса према протицају, укупан годишњи пронос суспендованог наноса као и специфична ерозија, утицај сезонских фактора на концентрацију силта и износ механичке ерозије, утицај геолошког састава на укупан пронос суспендованог наноса, мерење вученог наноса, однос хемијске и механичке ерозије. Хомоље и Звижд највећим делом припадају брдско – планинском подручју које се одликује водотоковима са релативно великим уздужним падовима, где су углови нагиба од 10 до 40. у Хомољу и Звижду све водотокове карактеришу врло велике разлике између екстремно великих и малих вода, па на неким водотоковима однос максималних и минималних протицаја достиже вредност и хиљаду. Овакав однос између малих и великих вода погодује развоју ерозивних процеса и разарању земљишта. Ерозија у сливу која почиње плувијалном ерозијом, а наставља се денудацијом, доводи до снижавања топографске површине и еродирања  земљишта. Таложењем наноса на дну речног корита оно се издиже, што доводи до смањивања овлаженог профила. Велике поплавне воде због тога не могу да протекну тако оплићалим речним коритом, па због тога долази до њиховог изливања, што се манифестује поплавама. Без земљишта остају горњи делови слива, а поплавама, услед таложења вученог и суспендованог наноса на дну речног корита, бивају изложени доњи делови речних токова. Овакав негативан тренд уочен је и у Хомољу и у Звижду који су последњих година све чешће угрожени поплавним водама које причињавају велике штете како пољопривредном земљишту тако и насељима у сливовима Млаве и Пека. С обзиром на чињеницу да у Хомољу и Звижду постоји велики број бујичних токова који носе огромне количине наноса, реално је очекивање да ће се тај нанос акумулирати у средњем и доњем делу сливова Млаве и Пека. Избор хидротехничког метода за истраживање обима и интензитета ерозије у Хомољу и Звижду, као и коришћење методе професора С. Гавриловића у анализи бујичних токова је покушај прецизнијег утврђивања квантитета овог процеса који последњих година има узлазни тренд. Остварење постављених задатака утврђеном методологијом и постигнуће задатих циљева пружа одговор на питање колики је интензитет и обим процеса ерозије у Хомољу и Звижду на годишњем нивоу, по годишњим добима и месецима. Утврђен је степен деловања физичко – географских фактора који имају утицај на процес ерозије, указано је на обим и врсту примене антиерозивних мера, нарочито у сливовима бујичних токова. Сагледавањем интензитета и обима ерозије у Хомољу и Звижду, односно сливовима Млаве и Пека, успешније ће бити решени водопривредни проблеми који овде постоје. 
Erozija zemljišta je jedan od najkompleksnijih problema i ujedno jedan od najpodmuklijih čovekovih neprijatelja. Erozivni procesi postepeno razaraju i uništavaju zemljište, odnose hranljive supstrate i remete vodni režim. Oni dovode stanovništvo mnogih, često prostranih oblasti u bedu, siromaštvo i migraciju. Zbog toga, mnoge privredne grane moraju biti uvučene u borbu protiv erozije zemljišta i njenih strahovitih posledica. Erozija je veoma složena dinamička pojava, koja zavisi od niza faktora čije je diferenciranje vrlo teško. Ispitivanje i utvrđivanje intenziteta vodne erozije spada u grupu problema koji se ne mogu sveobuhvatno rešiti. Veoma je teško, skoro nemoguće, izdvojiti sve faktore, a potom multiplicirati sve te faktore u svim postojećim kombinacijama. Ovim radom sagledano je stanje erozije zemljišta u slivu Mlave i Peka, odnosno Homolju i Zviždu, sa više aspekata i utvrđen niz pojava kao što su: koncentracija rastvorenih mineralnih materija, procentualno učešće makroelemenata u ukupnoj mineralizaciji, odnos mineralizacije prema proticaju, ukupan godišnji pronos rastvorenih mineralnih materija, odnos specifične električne provodljivosti i ukupne mineralizacije, iznos hemijske evakuacije, iznos hemijski rastvorenih mineralnih materija koje potiču od stenskog komleksa u slivu, uticaj geološkog sastava na ukupnu mineralizaciju vode, njenu strukturu i specifičnu hemijsku eroziju, koncentracija suspendovanog nanosa po profilima, odnos suspendovanog nanosa prema proticaju, ukupan godišnji pronos suspendovanog nanosa kao i specifična erozija, uticaj sezonskih faktora na koncentraciju silta i iznos mehaničke erozije, uticaj geološkog sastava na ukupan pronos suspendovanog nanosa, merenje vučenog nanosa, odnos hemijske i mehaničke erozije. Homolje i Zvižd najvećim delom pripadaju brdsko – planinskom području koje se odlikuje vodotokovima sa relativno velikim uzdužnim padovima, gde su uglovi nagiba od 10 do 40. u Homolju i Zviždu sve vodotokove karakterišu vrlo velike razlike između ekstremno velikih i malih voda, pa na nekim vodotokovima odnos maksimalnih i minimalnih proticaja dostiže vrednost i hiljadu. Ovakav odnos između malih i velikih voda pogoduje razvoju erozivnih procesa i razaranju zemljišta. Erozija u slivu koja počinje pluvijalnom erozijom, a nastavlja se denudacijom, dovodi do snižavanja topografske površine i erodiranja  zemljišta. Taloženjem nanosa na dnu rečnog korita ono se izdiže, što dovodi do smanjivanja ovlaženog profila. Velike poplavne vode zbog toga ne mogu da proteknu tako oplićalim rečnim koritom, pa zbog toga dolazi do njihovog izlivanja, što se manifestuje poplavama. Bez zemljišta ostaju gornji delovi sliva, a poplavama, usled taloženja vučenog i suspendovanog nanosa na dnu rečnog korita, bivaju izloženi donji delovi rečnih tokova. Ovakav negativan trend uočen je i u Homolju i u Zviždu koji su poslednjih godina sve češće ugroženi poplavnim vodama koje pričinjavaju velike štete kako poljoprivrednom zemljištu tako i naseljima u slivovima Mlave i Peka. S obzirom na činjenicu da u Homolju i Zviždu postoji veliki broj bujičnih tokova koji nose ogromne količine nanosa, realno je očekivanje da će se taj nanos akumulirati u srednjem i donjem delu slivova Mlave i Peka. Izbor hidrotehničkog metoda za istraživanje obima i intenziteta erozije u Homolju i Zviždu, kao i korišćenje metode profesora S. Gavrilovića u analizi bujičnih tokova je pokušaj preciznijeg utvrđivanja kvantiteta ovog procesa koji poslednjih godina ima uzlazni trend. Ostvarenje postavljenih zadataka utvrđenom metodologijom i postignuće zadatih ciljeva pruža odgovor na pitanje koliki je intenzitet i obim procesa erozije u Homolju i Zviždu na godišnjem nivou, po godišnjim dobima i mesecima. Utvrđen je stepen delovanja fizičko – geografskih faktora koji imaju uticaj na proces erozije, ukazano je na obim i vrstu primene antierozivnih mera, naročito u slivovima bujičnih tokova. Sagledavanjem intenziteta i obima erozije u Homolju i Zviždu, odnosno slivovima Mlave i Peka, uspešnije će biti rešeni vodoprivredni problemi koji ovde postoje. 
The abstract has been processed with OCR optical text recognition technology.Soil erosion is one of the most complex problems and at the same time one of the most insidious of human enemies. Erosion processes gradually destroy the land, carry away nutritional substrates and disrupt the water regime. They drive many people, often from large areas, to misery and poverty, and cause migration. Therefore, many industries have to be included into the fight against soil erosion and its terrible consequences. Erosion is a very complex and dynamic phenomenon that relies on many factors which are hard to define. Differentiating and determining the intensity of water erosion are among the problems that cannot be comprehensively solved. It is very difficult, almost impossible, to extract all the factors and then multiply all these factors in every possible combination. This paper looks at the situation with soil erosion in the water basins of the rivers Mlava and Pek respectively, that is, in the areas of Homolje and Zvižd, from various aspects, and identifies a series of phenomena such as concentration of dissolved minerals, percentage of macroelements in the total mineralization, the relationship of mineralization to the flow, the total annual transport of dissolved minerals, the ratio of the specific electrical conductivity and total mineralization, the amount of chemical evacuation, the amount of chemically dissolved minerals that come from the rock complex in the basin, the impact of geological composition on the total mineralization of water, its structure and specific chemical erosion, suspended sediment concentration by profiles, the effect of suspended sediment on the flow, the total annual transport of suspended sediment as well as the specific erosion, the impact of seasonal factors on the concentration of silt and the amount of mechanical erosion, the impact of geological composition on the total transport of suspended sediment, bed load measurement, the ratio of chemical and mechanical erosion. Homolje and Zvižd are mostly hilly and mountainous areas characterized by water streams with relatively large vertical drops where the pitch angles may vary from 10° to 40°. In Homolje and Zvižd all the waterways are characterized by very large differences between extremely high and low water, reaching the value of a thousand in some points. This kind of a ratio between high and low waters favors the development of erosion processes and the destruction of land. Erosion in a basin starts with a pluvial erosion and continues with denudation, causing the lowering of the topographic surface and soil erosion. When the sediment deposits at the bottom of the river bed, the bed rises, leading to a reduction of the humidified profile. Therefore great floodwater cannot pass through such a shallow riverbed, which leads to its discharge and causes a flood. The upper parts of the basin lose land, and the lower parts of the basin become exposed to floods due to the deposition of bed load and suspended sediment at the bottom of the riverbed. This negative trend has been observed both in Homolje and Zviždi which in recent years have become increasingly threatened by flood waters which cause great damage to agricultural land and settlements in the basins of the rivers Mlava and Pek. Considering the fact that Homolje and Zvižd there is a large number of torrential streams that carry huge amounts of sediment, it is reasonable to expect the sediment to accumulate in the middle and lower parts of the basins of Mlava and Pek. The selection of the hydrotechnical method to explore the scope and intensity of erosion in Homolje and Zvižd, and the use of Professor S. Gavrilović’s method in the analysis of torrential flows is an attempt to determine more precisely the quantity of this process with an increasing trend in recent years. Fulfillment of tasks and achievement of objectives through suggested methodology and achievement of objectives provides the answer to the question of intensity and extent of erosion in Homolje and Zvižd on an annual basis, according to seasons and months. The paper has determined the degree of impact of physical and geographic factors on the erosion process, and has pointed to the scope and type of application anti-erosion measures, especially in watersheds with torrential flows. If we manage to comprehend the intensity and extent of erosion in Homolje and Zvižd, that is, in the basins of Mlava and Pek, we will be able to solve the existing water management problems more successfully.

Recent years have seen the emergence of smart cities, which utilize various sensing data for applications such as pollution monitoring, infrastructure planning and traffic control. Current sensing projects tend to deploy a large number of low-cost and unreliable sensing sources, rather than a small number of high-quality sensing sources. It is therefore critical to provide data analysis in the face of unreliable sources. This thesis focuses on two types of sensing sources that have been used in smart city sensing projects, namely, environmental sensors and human sensors. The environmental sensors are physical sensors that are made to monitor certain environmental features, such as temperature, humidity, and pollutant concentration. An environmental sensor can fail frequently and will start generating faulty data when there is chemical compound decay, battery exhaustion, or calibration problems. Human sensors, as recently proposed in a new area called social sensing, are online messaging platform users who post observations about their surrounding environments. The data generated by human sensors can be erroneous because the natural language used in their messages does not conform to a machine-readable standard. Based on a survey of existing literature, this thesis presents source profiling-based solutions for three data analysis problems, data cleaning in environmental sensing, observation message classification in social sensing, and message location inference. Each of the solutions is validated with various real-world data and extensive experiments. For data cleaning in environmental sensing, we propose two solutions, approaching from a frequentist perspective and a Bayesian perspective, respectively. The frequentist approach determines sensor reliability based on the frequency of reliable behavior in the past, and in each data collection iteration updates a reliability score, which can be used to weight down or remove the data from unreliable sources. The Bayesian approach models sensor reliability as a latent variable, and applies the Expectation Maximization framework to discover the latent sensor reliability and correct reading values for the environmental feature. For observation message classification, we propose supervised and unsupervised solutions. We propose a supervised solution to distinguish messages according to three perspectives, namely, observation, affection, and speculation. We next propose a supervised solution based on user features such as trending activity, communication status, and writing styles. And finally, we propose an unsupervised solution based on lexical analysis and user profiling in four user attributes, namely, originality, interactivity, objectivity, and topic focus. For location inference, we propose a solution based on name entity extraction and user message histories. The proposed solution extracts location names from text messages using a gazetteer, and after retrieving a number of past locations from the message history of a user, it applies outlier removal before inferring the current location. Incorporating observation classification and location inference, we propose an event detection system called Sense and Focus (SNAF), which detects real world events based on discussions exchanged on Twitter. A prototype implementation of the system has shown a number of detection results, 54% of which corresponding to real-world events, and in many case detected earlier than news reports, and with less than 1.5km location error.
Thesis (Ph.D.) -- University of Adelaide, School of Computer Science, 2016.

碩士
國立交通大學
資訊學院資訊學程
100
Energy is always a critical resource in battery-driven devices. How hardware components are controlled and used by the applications and system software can significantly affect system’s energy consumption. In this thesis, we propose a measurement based energy profiling tool to assist software designer to determine different software design choices. The measurement based tool records target device’s power samples and system activities simultaneously, and correlating them together to generate device’s energy consumption report. Previous researches use statistical sampling or manual instrument method to record device’s system activities which may increase system overhead and lack of flexibility. We adopt dynamic source instrument technique and provide a GUI front-end tool to profile on the native (non-Java) part of Android system. The fine-grained process-level, function-level and code block-level performance and energy profiling reports are provided. The experiment results show the correctness of proposed profiling tool with low instrument overhead (0.02 milliseconds per function calls).

A World Wide Web e os seus recursos textuais crescem a cada dia que passa e a informação que pode ser extraída do comportamento dos utilizadores nas suas escolhas de conteúdos pode fazer crescer negócios e conduzir a novas ferramentas. Empresas como o Facebook ou o Google testemunham o que os utilizadores usam e partilham para fazer estudos do que pode vir a ser as necessidades futuras ou apenas construir o perfil de possíveis clientes. Conceitos de Data Mining como classificação ou associação podem ser usados para segmentar esses markets alcançando mais necessidades dos utilizadores ou obtendo a melhor campanha de marketing. Neste trabalho fazemos uma pequena pesquisa de como se pode fazer esse perfil e a sugestão de novos itens. Apresentamos uma ideianova, que pode ser usada para construção de perfis e sugestão de conteúdos que pode ser usada em grandes empresas como as mencionadas atrás.
The World Wide Web and its textual resources are growing every day and in-formation can be extracted from the behavior of users and what they are choos-ing to do in the Web can grow businesses and deliver new tools. Enterprises like Facebook or Google look to what users are using and sharing to conduct surveys of what could be the future demandsor even just profiling possible clients. Data Mining concepts such as classification or association can be used to segment such markets reaching more users’ needs or getting the perfect campaign done. In this work we conduct a small survey of what could be applied for profiling users and suggestion of new items. We present a novel idea that could be used in profiling and suggestion of contents that could be used in large enterprises like the ones mentioned above.

The hydrogeology, stable isotope distribution, and chemical distribution of Cl- and NO3--N within the Battersea Drainage Basin in southern Alberta were investigated. The Battersea Drainage Basin is characterized by widespread spreading of livestock manure on irrigated farmland and a high density of feedlots, creating concern about groundwater quality in the region. Past research has used conventional piezometers to study the source, distribution and fate of nitrate in the shallow groundwater. The key component of this research involved using the new technique of high-resolution profiling to determine the distribution, source, fate, and transport of nitrate in the shallow geological groundwater environment. High-resolution profiles of δ2H indicated groundwater throughout the glaciofluvial deposits and between 5.4 and 13.7 m below ground in glaciolacustrine deposits contained values > -150.0‰ and tritiated waters (> 0.08 TU). This suggested that this water recharged within the past 60 years. At depth 5.4 to 13.5 m BG, lower δ2H values did not coincide with detectable tritium, indicating the groundwater was much older and not vulnerable to agricultural contamination. High-resolution profiles of Cl- and NO3--N (solid core and squeezed pore water data) showed concentrations of these ions up to 411 and 219 mg L-1, respectively, in the glaciofluvial deposits. Concentrations of Cl- and NO3--N decreased to less than 50 mg L-1 (Cl-) and the Drinking Water Standard for NO3--N (10 mg L-1) in the underlying glaciolacustrine and glacial till deposits at most sites. Comparison to the high-resolution δ2H profiles suggested the high nitrate concentration in the glaciofluvial sediments is agricultural in origin. High concentrations for Cl- and NO3--N (up to 257 and 209 mg L-1, respectively) observed in glacial till and glaciolacustrine deposits below 6.0 m BG at two sites (LB5a and LB6) did not coincide with modern water, indicating the source of high nitrate is geologic in origin. The NO3--N to Cl- ratios suggested denitrification was not appreciable in the glaciofluvial deposits. However, denitrification may be a cause of decreased nitrate in the underlying fine textured deposits at certain sites. Interpretation of the high resolution profiles also suggested that the major conduit for nitrate migration is in near-surface glaciofluvial sediments via advection. The distributions of δ2H, Cl- and NO3--N with depth suggest that transport in the underlying glaciolacustrine and glacial till deposits is diffusion dominated, thus acting as a sink and removing nitrate from the permeable zone. However, the presence of fracturing in the oxidized zone of glacial tills and glaciolacustrine deposits suggests that transport may occur via advection through fracturing and diffusion in the material matrix. High-resolution profiling of δ2H, Cl- and NO3--N proved to be valuable in obtaining information regarding the distribution, source, and transport of groundwater and nitrate in the shallow groundwater environment at a level of detail that cannot be readily obtained through use of conventional piezometers.

M.Sc.
Drinking water treatment plants in South Africa rely almost entirely on surface water sources, which are often compromised due to high return flows and indirect reuse. The typical treatment plants focus on the removal of physical and microbial contaminants which include turbidity, colour, chemical compounds and micro-organisms. A relatively new alarm to this list is natural organic matter (NOM) which has become a major concern in potable water treatment due to its recent regulation. Conventionally, the drive to remove NOM from potable water would be the desire to remove colour from public water supplies. However, more problems in drinking water treatment associated with NOM have been recently identified. These include taste and odour, its tendency to foul membranes, interference with the removal of other contaminants and its potential to contribute to corrosion and slime growth in distribution systems. Moreover, it causes high demands for coagulants and disinfectants. The NOM is also the main precursor for disinfection by-products (DBPs) formed when it interacts with disinfectants such as chlorine during water disinfection. In this study, different raw water samples, of different NOM composition were collected from selected sources across the country and assessed for the removal of dissolved organic carbon (DOC) and UV absorbance at a wavelength of 254 nm (UV 254) using enhanced coagulation (EC). The efficacy of EC, which can be employed as a practical technology in the removal of both turbidity and NOM, was evaluated in these raw water sources. Jar tests were conducted with ferric chloride as the coagulant, and specific pH values were chosen as target values guiding the different coagulant dosages for the jar tests. The pH of the low-alkalinity (<60 mg/ℓ CaCO3) raw waters were adjusted and raised by the addition of sodium carbonate. Algorithms for finding the optimum coagulant dosage for both turbidity and UV 254 removal were developed and consistently applied to all the results in independent v batch tests, in which residual amounts of UV 254, DOC and turbidity were measured. The raw water parameters and results obtained from these tests were used to develop feed-forward multiplicative models predicting the performance of EC using ferric chloride. The results showed that the raw waters chosen were, indeed, representative of the different water types present in South Africa, and that the general water characteristics were affected by seasonal variations. The EC procedure developed was able to reduce turbidity to levels low enough for removal by subsequent treatment steps in the water treatment train. The residual UV 254 values were in all cases lower than 6 m-1, which theoretically corresponds to about 3.5 mg/ℓ DOC. This was confirmed by the residual DOC values which were found to be lower than 4 mg/ℓ. Generally, the waters of South Africa were found to be amenable to coagulation. In almost all cases, the costs for EC were comparable to those for conventional coagulation, hence EC could be employed as a NOM removal strategy in the South African context.

This research looks at the best practices applied by forensic investigators in conducting lifestyle audits on white collar crime suspects. The researcher explored, firstly, how a lifestyle audit relates to white collar crime investigations; secondly, the best practices in performing lifestyle audits on white collar crime suspects, and lastly, the various sources of information available to forensic investigators when conducting a lifestyle audit of a white collar crime suspect. As lifestyle audits will serve as an investigative tool in future forensic investigations into white collar crime, this study’s aim was to understand and create an awareness of the current best practices applied by forensic investigators within private sector forensic investigation practices while conducting lifestyle audits during white collar crime investigations. It also makes available research data regarding the concept of lifestyle audits in white collar crime investigations, the implementation of these audits, the benefits, and the best practices of these audits. The general purpose of this study was to provide practical recommendations on the best practices for lifestyle audits for forensic investigators within private sector.
Criminology
M. Tech. (Forensic Investigation)