Toward Integrating Feature Selection Algorithms for Classification and Clustering
IEEE Transactions on Knowledge and Data Engineering
A novel feature selection algorithm for text categorization
Expert Systems with Applications: An International Journal
Hi-index | 0.00 |
The feature reduction is one of the core techniques in text categorization. But there is no consideration of text position factor to the differentiation of labeling text capability in the method of weighting basing on multi-information (MI) in features. So in this paper, we put forward an improved feature selection method that based on MI. By adding the amending parameters in different positions, we have increased the using efficiency about the character information. The result of experiment shows that this method has improved the accuracy of the text classification.