Efficient text classification method based on improved term reduction and term weighting

Text classification has gained booming interest over the past few years. The traditional approaches of text classification commonly extract features from a signal test criterion, resulting in the problem of “over fitting”. This paper takes test criterions such as frequency, dispersion and concentration indices into account and proposes an improved dimension reduction method and feature weighting method, making the selection more representative and the weighting of characteristic features more reasonable. Experimental results show that the new method has high precision and recall rates.