Made available in DSpace on 2016-06-02T20:06:12Z (GMT). No. of bitstreams: 1 DissDAZ.pdf: 2962567 bytes, checksum: 5c6271a67fae12d6b0160ac8ed9351a2 (MD5) Previous issue date: 2006-02-20 / Universidade Federal de Minas Gerais / (See full text for download) / A crescente necessidade do desenvolvimento de eficientes técnicas computacionais e estatísticas para analisar a profusão de dados biológicos transformaram o modelo Markoviano oculto (HMM), caso particular das redes bayesianas ou probabilísticas, em uma alternativa interessante para analisar sequências de DNA. Uma razão do interesse no HMM é a sua flexibilidade em descrever segmentos heterogêneos da sequência através de uma mesma estrutura de dependência entre as variáveis, supostamente conhecida. No entanto, na maioria dos problemas práticos, a estrutura de dependência não é conhecida e precisa ser também estimada. A maneira mais comum para estimação de estrutra de um HMM é o uso de métodos de seleção de modelos. Outra solução é a utilização de metodologias para estimação da estrutura de uma rede probabilística. Neste trabalho, propomos o HMM de segunda ordem e seus estimadores bayesianos, definimos o fator de Bayes e o DIC para seleção do HMM mais adequado a uma sequência específica, verificamos seus desempenhos e a performance da metodologia proposta por Friedman e Koller (2003) em conjunto de dados simulados e aplicamos estas metodologias em duas sequências de DNA: o intron 7 do gene a - fetoprotein dos cimpanzés e o genoma do parasita Bacteriophage lambda, para o qual o modelo de segunda ordem é mais adequado.

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - CAPES / This work is a study of the relationship between the intrinsic dimension of feature vectors applied to the classification of video signals in order to perform lip reading. In pattern recognition tasks, the extraction of relevant features is crucial for a good performance of the classifiers. The starting point of this work was the reproduction of the work of J.R. Movellan [1], which classifies lips gestures with HMM using only the video signal from the Tulips1 database. The database consists of videos of volunteers’ mouths while they utter the first 4 numerals in English. The original work uses feature vectors of high dimensionality in relation to the size of the database. Consequently, the adjustment of HMM classifiers has become problematic and the maximum accuracy was only 66.67%. Alternative strategies for feature extraction and classification schemes were proposed in order to analyze the influence of the intrinsic dimension in the performance of classifiers. The best solution, in terms of results, achieved an accuracy of approximately 83%. / Este trabalho é um estudo da relação entre a dimensão intrínseca de vetores de características aplicados à classificação de sinais de vídeo no intuito de realizar-se a leitura labial. Nas tarefas de reconhecimento de padrões, a extração de características relevantes é crucial para um bom desempenho dos classificadores. O ponto de partida deste trabalho foi a reprodução do trabalho de J.R. Movellan [1], que realiza a classificação de gestos labiais com HMM na base de dados Tulips1, utilizando somente o sinal de vídeo. A base é composta por vídeos das bocas de voluntários enquanto esses pronunciam os primeiros 4 numerais em inglês. O trabalho original utiliza vetores de características de dimensão muito alta em relação ao tamanho da base. Consequentemente, o ajuste de classificadores HMM se tornou problemático e só se alcançou 66,67% de acurácia. Estratégias de extração de características e esquemas de classificação alternativos foram propostos, a fim de analisar a influência da dimensão intrínseca no desempenho de classificadores. A melhor solução, em termos de resultados, obteve uma acurácia de aproximadamente 83%. / São Cristóvão, SE

Current Ambient Intelligence and Intelligent Environment research focuses on the interpretation of a subject’s behaviour at the activity level by logging the Activity of Daily Living (ADL) such as eating, cooking, etc. In general, the sensors employed (e.g. PIR sensors, contact sensors) provide low resolution information. Meanwhile, the expansion of ubiquitous computing allows researchers to gather additional information from different types of sensor which is possible to improve activity analysis. Based on the previous research about sitting posture detection, this research attempts to further analyses human sitting activity. The aim of this research is to use non-intrusive low cost pressure sensor embedded chair system to recognize a subject’s activity by using their detected postures. There are three steps for this research, the first step is to find a hardware solution for low cost sitting posture detection, second step is to find a suitable strategy of sitting posture detection and the last step is to correlate the time-ordered sitting posture sequences with sitting activity. The author initiated a prototype type of sensing system called IntelliChair for sitting posture detection. Two experiments are proceeded in order to determine the hardware architecture of IntelliChair system. The prototype looks at the sensor selection and integration of various sensor and indicates the best for a low cost, non-intrusive system. Subsequently, this research implements signal process theory to explore the frequency feature of sitting posture, for the purpose of determining a suitable sampling rate for IntelliChair system. For second and third step, ten subjects are recruited for the sitting posture data and sitting activity data collection. The former dataset is collected byasking subjects to perform certain pre-defined sitting postures on IntelliChair and it is used for posture recognition experiment. The latter dataset is collected by asking the subjects to perform their normal sitting activity routine on IntelliChair for four hours, and the dataset is used for activity modelling and recognition experiment. For the posture recognition experiment, two Support Vector Machine (SVM) based classifiers are trained (one for spine postures and the other one for leg postures), and their performance evaluated. Hidden Markov Model is utilized for sitting activity modelling and recognition in order to establish the selected sitting activities from sitting posture sequences.2. After experimenting with possible sensors, Force Sensing Resistor (FSR) is selected as the pressure sensing unit for IntelliChair. Eight FSRs are mounted on the seat and back of a chair to gather haptic (i.e., touch-based) posture information. Furthermore, the research explores the possibility of using alternative non-intrusive sensing technology (i.e. vision based Kinect Sensor from Microsoft) and find out the Kinect sensor is not reliable for sitting posture detection due to the joint drifting problem. A suitable sampling rate for IntelliChair is determined according to the experiment result which is 6 Hz. The posture classification performance shows that the SVM based classifier is robust to “familiar” subject data (accuracy is 99.8% with spine postures and 99.9% with leg postures). When dealing with “unfamiliar” subject data, the accuracy is 80.7% for spine posture classification and 42.3% for leg posture classification. The result of activity recognition achieves 41.27% accuracy among four selected activities (i.e. relax, play game, working with PC and watching video). The result of this thesis shows that different individual body characteristics and sitting habits influence both sitting posture and sitting activity recognition. In this case, it suggests that IntelliChair is suitable for individual usage but a training stage is required.

The vital signs monitoring data of an infant receiving intensive care are a rich source of information about its health condition. One major concern about the state of health of such patients is the onset of neonatal sepsis, a life-threatening bloodstream infection. As early signs are subtle and current diagnosis procedures involve slow laboratory testing, sepsis detection based on the monitored physiological dynamics is a clinically significant task. This challenging problem can be thoroughly modelled as real-time inference within a machine learning framework. In this thesis, we develop probabilistic dynamical models centred around the goal of providing useful predictions about the onset of neonatal sepsis. This research is characterised by the careful incorporation of domain knowledge for the purpose of extracting the infant’s true physiology from the monitoring data. We make two main contributions. The first one is the formulation of sepsis detection as learning and inference in an Auto-Regressive Hidden Markov Model (AR-HMM). The model investigates the extent to which physiological events observed in the patient’s monitoring traces could be used for the early detection of neonatal sepsis. In addition, the proposed approach involves exact marginalisation over missing data at inference time. When applying the ARHMM on a real-world dataset, we found that it can produce effective predictions about the onset of sepsis. Second, both sepsis and clinical event detection are formulated as learning and inference in a Hierarchical Switching Linear Dynamical System (HSLDS). The HSLDS models dynamical systems where complex interactions between modes of operation can be represented as a twolevel hidden discrete hierarchical structure. For neonatal condition monitoring, the lower layer models clinical events and is controlled by upper layer variables with semantics sepsis/nonsepsis. The model parameterisation and estimation procedures are adapted to the specifics of physiological monitoring data. We demonstrate that the performance of the HSLDS for the detection of sepsis is not statistically different from the AR-HMM, despite the fact that the latter model is given “ground truth” annotations of the patient’s physiology.

L'apnée est une complication fréquente chez les nouveaux-nés prématurés. L'un des problèmes les plus fréquents est l'épisode d'apnée bradycardie dont la répétition influence de manière négative le développement de l'enfant. C'est pourquoi les enfants prématurés sont surveillés en continu par un système de monitoring. Depuis la mise en place de ce système, l'espérance de vie et le pronostic de vie des prématurés ont été considérablement améliorés et ainsi la mortalité réduite. En effet, les avancées technologiques en électronique, informatique et télécommunications ont conduit à l'élaboration de systèmes multivoies de monitoring néonatal de plus en plus performants. L'un des principaux signaux exploités dans ces systèmes est l'électrocardiogramme (ECG). Toutefois, même si l'analyse de l'ECG a évolué au fil des années, l'ensemble des informations qu'il fournit n'est pas encore totalement exploité dans les processus de décision, notamment en monitoring en Unité de Soins Intensifs en Néonatalogie (USIN). L'objectif principal de cette thèse est d'améliorer la prise en compte des dynamiques multi-dimensionnelles en proposant de nouvelles approches basées sur un formalisme bayésien, pour la détection précoce des apnées bradycardies chez le nouveau-né prématuré. Aussi, dans cette thèse, nous proposons deux approches bayésiennes, basées sur les caractéristiques de signaux biologiques en vue de la détection précoce de l'apnée bradycardie des nouveaux-nés prématurés. Tout d'abord avec l'approche de Markov caché, nous proposons deux extensions du Modèle de Markov Caché (MMC) classique. La première, qui s'appelle Modèle de Markov Caché Couplé (MMCC), créé une chaîne de Markov à chaque dimension de l'observation et établit un couplage entre les chaînes. La seconde, qui s'appelle Modèle Semi-Markov Caché Couplé (MSMCC), combine les caractéristiques du modèle de MSMC avec le mécanisme de couplage entre canaux. Pour les deux nouveaux modèles (MMCC et MSMCC), les algorithmes récursifs basées sur la version classique de Forward-Backward sont introduits pour résoudre les problèmes d'apprentissage et d'inférence dans le cas couplé. En plus des modèles de Markov, nous proposons deux approches passées sur les filtres de Kalman pour la détection d'apnée. La première utilise les modifications de la morphologie du complexe QRS et est inspirée du modèle générateur de McSharry, déjà utilisé en couplant avec un filtre de Kalman étendu dans le but de détecter des changements subtils de l'ECG, échantillon par échantillon. La deuxième utilise deux modèles AR (l'un pour le processus normal et l'autre pour le processus de bradycardie). Les modèles AR sont appliqués sur la série RR, alors que le filtre de Kalman suit l'évolution des paramètres du modèle AR et fournit une mesure de probabilité des deux processus concurrents. / Apnea-bradycardia episodes (breathing pauses associated with a significant fall in heart rate) are the most common disease in preterm infants. Consequences associated with apnea-bradycardia episodes involve a compromise in oxygenation and tissue perfusion, a poor neuromotor prognosis at childhood and a predisposing factor to sudden-death syndrome in preterm newborns. It is therefore important that these episodes are recognized (early detected or predicted if possible), to start an appropriate treatment and to prevent the associated risks. In this thesis, we propose two Bayesian Network (BN) approaches (Markovian and Switching Kalman Filter) for the early detection of apnea bradycardia events on preterm infants, using different features extracted from electrocardiographic (ECG) recordings. Concerning the Markovian approach, we propose new frameworks for two generalizations of the classical Hidden Markov Model (HMM). The first framework, Coupled Hidden Markov Model (CHMM), is accomplished by assigning a Markov chain (channel) to each dimension of observation and establishing a coupling among channels. The second framework, Coupled Hidden semi Markov Model (CHMM), combines the characteristics of Hidden semi Markov Model (HSMM) with the above-mentioned coupling concept. For each framework, we present appropriate recursions in order to use modified Forward-Backward (FB) algorithms to solve the learning and inference problems. The proposed learning algorithm is based on Maximum Likelihood (ML) criteria. Moreover, we propose two new switching Kalman Filter (SKF) based algorithms, called wave-based and R-based, to present an index for bradycardia detection from ECG. The wave-based algorithm is established based on McSarry's dynamical model for ECG beat generation which is used in an Extended Kalman filter algorithm in order to detect subtle changes in ECG sample by sample. We also propose a new SKF algorithm to model normal beats and those with bradycardia by two different AR processes.

In Natural Language Processing (NLP), speech and text are parsed and generated with language models and parser models, and translated with translation models. Each model contains a set of numerical parameters which are found by applying a suitable training algorithm to a set of training data. Many such training algorithms are instances of the Expectation-Maximization (EM) algorithm. In [BSV15], a generic EM algorithm for NLP is described. This work presents a particular speech model, the Hidden Markov model, and its standard training algorithm, the Baum-Welch algorithm. It is then shown that the Baum-Welch algorithm is an instance of the generic EM algorithm introduced by [BSV15], from which follows that all statements about the generic EM algorithm also apply to the Baum-Welch algorithm, especially its correctness and convergence properties.

Les modèles latents sont très utilisés en pratique, comme en génomique, économétrie, reconnaissance de parole... Comme la modélisation paramétrique des densités d’émission, c’est-à-dire les lois d’une observation sachant l’état latent, peut conduire à de mauvais résultats en pratique, un récent intérêt pour les modèles latents non paramétriques est apparu dans les applications. Or ces modèles ont peu été étudiés en théorie. Dans cette thèse je me suis intéressée aux propriétés asymptotiques des estimateurs (dans le cas fréquentiste) et de la loi a posteriori (dans le cadre Bayésien) dans deux modèles latents particuliers : les modèles de Markov caché et les modèles de mélange. J’ai tout d’abord étudié la concentration de la loi a posteriori dans les modèles non paramétriques de Markov caché. Plus précisément, j’ai étudié la consistance puis la vitesse de concentration de la loi a posteriori. Enfin je me suis intéressée à l’estimation efficace du paramètre de mélange dans les modèles semi paramétriques de mélange. / Latent models have been widely used in diverse fields such as speech recognition, genomics, econometrics. Because parametric modeling of emission distributions, that is the distributions of an observation given the latent state, may lead to poor results in practice, in particular for clustering purposes, recent interest in using non parametric latent models appeared in applications. Yet little thoughts have been given to theory in this framework. During my PhD I have been interested in the asymptotic behaviour of estimators (in the frequentist case) and the posterior distribution (in the Bayesian case) in two particuliar non parametric latent models: hidden Markov models and mixture models. I have first studied the concentration of the posterior distribution in non parametric hidden Markov models. More precisely, I have considered posterior consistency and posterior concentration rates. Finally, I have been interested in efficient estimation of the mixture parameter in semi parametric mixture models.

In this thesis, methods of implementing an online gesture recognition system for the novel Stylaero Board device are investigated. Two methods are evaluated - one based on LSTMs and one based on HMMs - on three kinds of gestures: Tap, circle, and flick motions. A method’s performance was measured in its accuracy in determining both whether any of the above listed gestures were performed and, if so, which gesture, in an online single-pass scenario. Insight was acquired regarding the technical challenges and possible solutions to the online aspect of the problem. Poor performance was, however, observed in both methods, with a likely culprit identified as low quality of training data, due to an arduous and complex gesture performance capturing process. Further research improving on the process of gathering data is suggested.

Financial data is often said to follow different market regimes. These regimes, which not possible to observe directly, are assumed to influence the observable returns. In this thesis such regimes are modeled using hidden Markov models. We will investigate whether the five different currency pairs EUR/NOK, USD/NOK, EUR/USD, EUR/SEK, and USD/SEK exhibit market regimes that can be described using hidden Markov modeling. We will find the most optimal number of states and study the mean, variance, and correlations in each market regime. / Finansiella data sägs ofta följa olika marknadsregimer. Dessa marknadsregimer kan inte observeras direkt men antas ha inflytande på de observerade avkastningarna. I denna uppsats undersöks om de fem valutaparen EUR/NOK, USD/NOK, EUR/USD, EUR/SEK och USD/SEK tycks följa separata marknadsregimer som kan detekteras med hjälp av en dold Markovkedja.

Philosophiae Doctor - PhD / This chapter outlines basic gene structure and how gene structure is related to promoter structure in both prokaryotes and eukaryotes and their transcription machinery. An in-depth discussion is given on variations types of the promoters among both prokaryotes and eukaryotes and as well as among three prokaryotic organisms namely, E.coli, B.subtilis and Mycobacteria with emphasis on Mituberculosis. The simplest definition that can be given for a promoter is: It is a segment of Deoxyribonucleic Acid (DNA) sequence located upstream of the 5' end of the gene where the RNA Polymerase enzyme binds prior to transcription (synthesis of RNA chain representative of one strand of the duplex DNA). However, promoters are more complex than defined above. For example, not all sequences upstream of genes can function as promoters even though they may have features similar to some known promoters (from section 1.2). Promoters are therefore specific sections of DNA sequences that are also recognized by specific proteins and therefore differ from other sections of DNA sequences that are transcribed or translated. The information for directing RNA polymerase to the promoter has to be in section of DNA sequence defining the promoter region. Transcription in prokaryotes is initiated when the enzyme RNA polymerase forms a complex with sigma factors at the promoter site. Before transcription, RNA polymerase must form a tight complex with the sigma/transcription factor(s) (figure 1.1). The 'tight complex' is then converted into an 'open complex' by melting of a short region of DNA within the sequence involved in the complex formation. The final step in transcription initiation involves joining of first two nucleotides in a phosphodiester linkage (nascent RNA) followed by the release of sigma/transcription factors. RNA polymerase then continues with the transcription by making a transition from initiation to elongation of the nascent transcript.