Discussion About The Naive Bayesian Classifier Based On Selected And Weighted Attributes

The Naive Bayesian Classifier is a simple kind of Bayesian Classification Model. It has been widely applied in the field of decision making and classification because of its high accuracy and high efficiency, especially its solid theory foundation. But this classifier requires attributes are mutually independent with given feature of classification, meantime, it makes all condition attributes the same effect to the decision attributes. However, the real data don't usually meet these requirements which become the method's disadvantages.This paper introduces the Bayesian Decision Theory, the Naive Bayesian Classifier, the advantages and disadvantages of this classifier, in order to reduce its limitations, this paper introduces methods of attribute selection and weight, and proposes a new improved bayesian classification model applied to improve attributes selection and weight based on the method of General Correlation Function, the Classifier, that is, the SWGNBC. In the end, the experiment results on UCI datasets show that the improved model proposed in the paper has a better performance.However, there are still two points need future discussion and research. The one is whether the measurement based on General Correlation Function is the best for improving classification results. The other is whether the improved classifier can reach the same high classification accuracy on Continuous Datasets.