Systems, Man, and Cybernetics

 

Home

• [P803] Study on Feature Selection in Finance Text Categorization
  Author(s): Changqiu Sun
  
  

  Abstract
  This paper proposes a new feature selection method Opposite Words (OW). We make a comparative study and a combinative application with other eight feature selection methods in text categorization. Combined with the classic VSM classifier based on cosine similarity, SVM and the Na?ve Bayes classifier, training and testing are carried out on two text sets with different class distribution. As the results indicate, OW suits for the case that features are in a small amount, especially when corpus is abundant in opposite words. Besides, OW is a good teammate. When we make this freshman cooperate with other big brothers, they come near to perfection for features in general or large amount.

  
   

Full Text (PDF)