Revisiting a two-stage feature selection method for text classification
What is it about?
Recently, a two-stage feature selection method for text classification was introduced. The method combines class-based and corpus-based metrics. Based on their experiments, the authors conclude what parameter values for each stage, allow a feature selection which improves the traditional methods in text classification. In this work, we revisited this two-stage feature selection method and based on several experiments we found a new way to select the parameters, which provides better results than the original work.
Why is it important?
This work is important because a small improvement is obtained over a method proposed in the state of the art for feature selection in text classification domain. In addition, it is an example for any researcher, how a detailed review of an article can be converted into a new article with different results.