An Effective Feature Selection Method for Text Categorization

摘要: Feature selection is an efficient strategy to reduce the dimensionality of data and removing noise in text categorization. However, most feature methods aim remove non-informative features based on corpus statistics, which do not relate classification accuracy directly. In this paper, we propose effective method, aims at KNN. Our experiments show that our method better than traditional methods, it also beneficial other classifiers, such as Support Vector Machines (SVM).