一种基于独立类别特性的改进KNN文本分类算法

With the fast development of World Wide Web, text categorization in data mining has become the key technology in organizing and processing large amount of document data. K-Nearest Neighbor(KNN) algorithm, as a simple, effective and nonparametric classification method, has the advantage of high stability and accuracy and is widely used in various industry sectors. Especially in the field of information security, KNN has become one of the important algorithms to filter text content. Class imbalance is one of the problems plagueing practitioners in data mining community. When training set is skewed, the popular KNN text classifier will mislabel instances in rare categories into common ones and lead to degradation in macro F1. Many solutions have been proposed. In this paper, an improved KNN text categorization algorithm based on independent category feature is presented. The experiment shows that it has good performance.