The Research of Text Classification Technology Based on Improved Maximum Entropy Model

In this paper, it studies the Techniques of Text Categorization based on maximum entropy models, and gives technical solutions for the network through improved technology, which contains a large number of documents. This paper describes an method based on improved maximum entropy model for text classification, a features selection method and pretreatment methods, the experimental results with good accuracy and recall, has played a certain role for large amounts of text, to meet the requirements of practical application.