Stastical pattern recognition is the most successful approach to automatic speech and speaker recognition (ASASR). Of all the statistical pattern recognition techniques, the hidden Markov model (HMM) is the most important. The Gaussian mixture model (GMM) and vector quantisation (VQ) are also effective techniques, especially for speaker recognition and in conjunction with HMMs. for speech recognition. However, the performance of these techniques degrades rapidly in the context of insufficient training data and in the presence of noise or distortion. Fuzzy approaches with their adjustable parameters can reduce such degradation. Fuzzy set theory is one of the most, successful approaches in pattern recognition, where, based on the idea of a fuzzy membership function, fuzzy C'-means (FCM) clustering and noise clustering (NC) are the most, important techniques. To establish fuzzy approaches to ASASR, the following basic problems are solved. First, a time-dependent fuzzy membership function is defined for the HMM. Second, a general distance is proposed to obtain a relationship between modelling and clustering techniques. Third, fuzzy entropy (FE) clustering is proposed to relate fuzzy models to statistical models. Finally, fuzzy membership functions are proposed as discriminant functions in decison making. The following models are proposed: 1) the FE-HMM. NC-FE-HMM. FE-GMM. NC-FEGMM. FE-VQ and NC-FE-VQ in the FE approach. 2) the FCM-HMM. NC-FCM-HMM. FCM-GMM and NC-FCM-GMM in the FCM approach, and 3) the hard HMM and GMM as the special models of both FE and FCM approaches. Finally, a fuzzy approach to speaker verification and a further extension using possibility theory are also proposed. The evaluation experiments performed on the TI46, ANDOSL and YOHO corpora showbetter results for all of the proposed techniques in comparison with the non-fuzzy baseline techniques.

A utilização de metodologias Bayesianas tem se tornado freqüuente nas aplicações em Genética, em particular em mapeamento de QTLs usando marcadores moleculares. Mapear um QTL implica em identificar sua posição no genoma, bem como seus efeitos genéticos. A abordagem Bayesiana combina, através do Teorema de Bayes, a verossimilhança dos dados fenotípicos com distribuições a priori atribuídas a todos os parâmetros desconhecidos (número, localização e efeito do QTL) induzindo distribuições a posteriori a respeito dessas quantidades. Métodos de mapeamento Bayesiano podem tratar o número desconhecido de QTLs como uma variável aleatória, resultando em complicações na obtençãao da amostra aleatória da distribuição conjunta a posteriori, uma vez que a dimensão do espaço do modelo pode variar. O Método MCMC com Saltos Reversíveis (MCMC-SR), proposto por Green(1995), é excelente para explorar distribuições a posteriori nesse contexto. O método proposto foi avaliado usando dados simulados no WinQTLCart, onde o maior objetivo foi avaliar diferentes prioris atribuídas para o número de QTLs. / The use of Bayesian methodology in genetical applications has grown increasingly popular, in particular in the analysis of quantitative trait loci (QTL) for studies using molecular markers. In such analyses the aim is mapping QTLs, estimating their locations in the genome and their genotypic effects. The Bayesian approach proceeds by setting up a likelihood function for the phenotype and assigning prior distributions to all unknowns in the problem (number of QTL, chromosome, locus, genetics effects). These induce a posterior distribution on the unknown quantities that contains all of the available information for inference of the genetic architecture of the trait. Bayesian mapping methods can treat the unknown number of QTL as a random variable, which has several advantages but results in the complication of varying the dimension of the model space. The reversible jump MCMC algorithm offers a powerful and general approach to exploring posterior distributions in this setting. The method was evaluated by analyzing simulated data, where the major goal was evaluate if different priors distributions on the QTL numbers.

A utilização de metodologias Bayesianas tem se tornado freqüuente nas aplicações em Genética, em particular em mapeamento de QTLs usando marcadores moleculares. Mapear um QTL implica em identificar sua posição no genoma, bem como seus efeitos genéticos. A abordagem Bayesiana combina, através do Teorema de Bayes, a verossimilhança dos dados fenotípicos com distribuições a priori atribuídas a todos os parâmetros desconhecidos (número, localização e efeito do QTL) induzindo distribuições a posteriori a respeito dessas quantidades. Métodos de mapeamento Bayesiano podem tratar o número desconhecido de QTLs como uma variável aleatória, resultando em complicações na obtençãao da amostra aleatória da distribuição conjunta a posteriori, uma vez que a dimensão do espaço do modelo pode variar. O Método MCMC com Saltos Reversíveis (MCMC-SR), proposto por Green(1995), é excelente para explorar distribuições a posteriori nesse contexto. O método proposto foi avaliado usando dados simulados no WinQTLCart, onde o maior objetivo foi avaliar diferentes prioris atribuídas para o número de QTLs. / The use of Bayesian methodology in genetical applications has grown increasingly popular, in particular in the analysis of quantitative trait loci (QTL) for studies using molecular markers. In such analyses the aim is mapping QTLs, estimating their locations in the genome and their genotypic effects. The Bayesian approach proceeds by setting up a likelihood function for the phenotype and assigning prior distributions to all unknowns in the problem (number of QTL, chromosome, locus, genetics effects). These induce a posterior distribution on the unknown quantities that contains all of the available information for inference of the genetic architecture of the trait. Bayesian mapping methods can treat the unknown number of QTL as a random variable, which has several advantages but results in the complication of varying the dimension of the model space. The reversible jump MCMC algorithm offers a powerful and general approach to exploring posterior distributions in this setting. The method was evaluated by analyzing simulated data, where the major goal was evaluate if different priors distributions on the QTL numbers.

The thesis is dedicated to the topic of employment and motivation of workers in tourism segment. The goal is to get a structure of motivational factors that are important for workers in travel agencies, hotels and information centres and find out, how the factors are present in the scope of employment. On this research the research among students of tourism schools takes up. The results are compared to the similar researches made on the Czech employees as a whole. For better understanding of the importance of workers for tourism business the thesis contains basic overview of the business that serve the tourist participants and detailed overview of the sources of statistical data regarding employment in the tourism segment.