Feature Selection Methods Research in Naive Bayes Spam Filtering Model
Jiaoyan Zhang · Aeronautical Computing Technique · 2008
The feature selection methods of Text Categorization are introduced into Naive Bayes Filtering(NBF) model in Spam Filtering,including document frequency,information gain,mutual information,expected cross entropy,weight of evidence for text,odds ratio,and x2 statistics(CHI).Characteristic of these methods are evaluated and analyzed.The test results demonstrate that x2 statistics is superior to other methods.