Incremental Bayes Algorithm Based On Niche Genetic Algorithm

With the development of digital technologies, the amount of data was enlarged fast. There was, however, some important information contained by the mass data. Therefore, people hoped that the mass data could be utilized to process weather report, expert diagnosis and so on through analyzing the mass data. Data classification came out for that.Na(?)ve Bayes algorithm based on two assumptions was used popularly. One was attribute independence. The other was Bayes' theorem. The most phenomenal advantage of Na(?)ve Bayes algorithm was to compute the approximate value of probability of each attribution through calculating the occurrence probability of each attribution of training examples, which was not to search the whole example space. Nevertheless, there were still some problems. First, Na(?)ve Bayes algorithm could not process the problem of incremental classification. Second, the computation was complex when it processed the classification. Third, it could not make full use of information from the classification at the first time.According to three problems mentioned above, incremental Bayes algorithm based on Niche Genetic Algorithm was presented with importing the conception of vector space. With the quantizing example via vector space, Niche Genetic Algorithm was used to extract features of examples in local spaces as the standards to classify, which improved the efficiency and the precision. Finally, two experiments were accomplished to illustrate that the time complexity and the space complexity were brought down, and the algorithm could be utilized to process the incremental classification.