Improveing Based On Naive Bayesian Classifier Algorithm

Data Mining is a technique in the field of machine learning. It closely combines artificial intelligent technique with data base technique, which can make computers help collect valuable knowledge models intelligently and automatically from large dada to satisfy people's different practical needs. Bayesian classification model is one of algorithm of data mining. Based on Bayesian theory, utmost posteriori hypothesis, Bayesian network and information science theories, this algorithm, using transcendence statistic knowledge owns perfect characteristics which can easy to calculate, and effectively deal with classification problems of mixed index attribute and it has sound classification competence. So it has developed at present as one of data mining theory and applicable research method.This paper mainly discusses about the research and application of Bayesian classification model.Firstly, it gives a general introduction about basic theory of Bayesian classification.Secondly, it concludes several frequently-used Bayesian classification models, which are naive Bayesian classification model, CB classification model, classification models based on relevance attribute, EANBC classification model, na?ve Bayesian classification model based on feature weighting and Bayesian classification model limited on attribute.Thirdly, it focuses on improvement of na?ve Bayesian classification calculation. It puts structure thought ofχ2statistic volume into relevant measure formula with certain structure attribute to improve Bayesian classification calculation. Furthermore, it proves by experiment that the improved calculation mentioned in this paper can effectively enhance the classification effect.Lastly, it does a conclusion for this paper and put forward some suggestions about further study.