The design of an optimal Bayesian classifier for multiple features is dependent on the estimation of multidimensional joint probability density functions and therefore requires a design sample size that increases exponentially with the number of dimensions. A method was developed that combines classifications from marginal density functions using an additional classifier. Unlike voting methods,...

The ensembles of simple Bayesian classifiers have traditionally not been a focus of research. The reason is that simple Bayes is an extremely stable learning algorithm and most ensemble techniques such as bagging is mainly variance reduction techniques, thus not being able to benefit from its integration. However, simple Bayes can be effectively used in ensemble techniques, which perform also b...

In the emerging field of toxicogenomics, microarray technology is used to read the level of gene expression in cells subjected to various toxic and non toxic agents. The amount of levels that are read from one single array is in the thousands, and when the number of arrays increases, statistical methods are needed to process the information. This master’s thesis documents the use of Bayesian st...

This paper presents a number of proofs that equate the outputs of a Multi-Layer Perceptron (MLP) classifier and the optimal Bayesian discriminant function for asymptotically large sets of statistically independent training samples. Two broad classes of objective functions are shown to yield Bayesian discriminant performance. The first class are “reasonable error measures,” which achieve Bayesia...

In multi-label classification the goal is to assign an instance to a set of different classes. This task is normally addressed either by defining a compound class variable with all the possible combinations of labels (label power-set methods) or by building independent classifiers for each class (binary relevance methods). The first approach suffers from high computationally complexity, while t...

A novel Bayesian classifier with smaller eigenvalues reset by threshold based on database is proposed in this paper. The threshold is used to substitute eigenvalues of scatter matrices which are smaller than the threshold to minimize the classification error rate with a given database, thus improving the performance of Bayesian classifier. Several experiments have shown its effectiveness. The e...

The Naive Bayesian classifier (NBC) is a well-known classification model that has simple structure, low training complexity, excellent scalability, and good performances. However, the NBC two key limitations: (1) it built upon strong assumption condition attributes are independent, which often does not hold in real-life, (2) handle continuous well. To overcome these limitations, this paper pres...