A Text Classification Method Based on Cascade

Abstract

Text categorization is an important research area in many Information Retrieval applications. Based on the research of text classification technology, a new bi-class text categorization method, which uses a cascade of AdaBoost classifiers, is proposed. Experiment on Fudan University data set shows that this method has a good classify performance at 92.167{%}. Compared with other classification methods, it has a lower storage space, smaller computation time and is suitable for different situations.