A correlational Bayesian network for DNA microarray data analysis

From DNA microarray experiments, we need to deal with large datasets in which instances are described by many features. In order to reduce high data dimensionality, feature selection method (FSMs) is usually applied in the flow of data analysis. In this paper, we propose a correlational Bayesian network for feature selection. The proposed algorithm is able to effectively identify and manipulate correlational individuals so that it improves performance and provides higher accurate results than other Bayesian network learning and FSMs. Through use of Bayesian framework to infer the weights, weight decay terms and perform model selection, we can obtain neural models with high generalization capability and low complexity. As a classifier Backpropagation network is used for classification of cancer types. The experiments are carried out for verification of the proposed method. A comparison study is also done with conventional Bayesian network approach and other FSMs. From comparison it can be seen that the correlational Bayesian network (CBN) proposed in thia paper is effective.