Naive Bayesian classifier based on genetic simulated annealing algorithm

摘要

Naive Bayesian classifier (NBC) is a simple and effective classifier, but in the actual application, its attribute independence assumption is not always set up. This factor affects its classification performance. Attribute reduction is an effective way to improve the performance of this classification. This paper take advantage of mixed Simulated Annealing and Genetic Algorithms to optimize attribute set, so that a better NBC is constructed. Comparing the based on genetic algorithm NBC with the traditional NBC, experiment results show that the based on genetic algorithm NBC can be more effective and rapid to solve the classification performance of NBC.