An Improved Random Forest Classifier for Text Categorization

This paper proposes an improved random forest algorithm for classifying text data. This algorithm is particularly designed for analyzing very high dimensional data with multiple classes whose well-known representative data is text corpus. A novel feature weighting method and tree selection method are developed and synergistically served for making random forest framework well suited to categorize text documents with dozens of topics. With the new feature weighting method for subspace sampling and tree selection method, we can effectively reduce subspace size and improve classification performance without increasing error bound. We apply the proposed method on six text data sets with diverse characteristics. The results have demonstrated that this improved random forests outperformed the popular text classification methods in terms of classification performance. Index Terms — random forest, text categorization, random subspace, decision tree