论文部分内容阅读
Calculating the weight of feature words is a key step in text classification.A method,based on the traditional method—TFIDF,coupled with feature entropy,introducing minimum classification error to train a parameter of every word is proposed,and then is applied for spam sms filtering.