Research And Implementation Of Text Categorization System Using Feedback Methods Based On VSM

In recent years, information categorization turns more and more important for us to get useful information . Text Categorization, i.e. automated assigning texts to predefined categories based on their contents, is a task of increasing importance.Now, Vector Space Model (VSM) is the best model for large scale of text processing. Firstly, We discuss the key techniques of VSM, including: basic conception of VSM, Feature Selection and Feature Extraction.The second part is the introduction to several common Text Categorization methods and the algorithms are presented in detail.In many important text classification problems, acquiring class labels for training documents is costly. This paper show that the accuracy of text classifiers trained with a small number of labeled documents can be improved by using feedback methods. The proposed classification system is divided into three parts: training procedure, classifying procedure and feedback procedure. The system has good scalability and flexility. Based on the text classification system, we have done much work on testing and have got much precise data. All these data show that under conditions of inadequate training, the text classification system using feedback methods can achieve good performance.