网页分类中特征选择方法的研究

Research of feature selection methods of web page classification system

准确的网络分类对于健康的网络环境至关重要,本文基于这样的目的,采用了效果理想SVM分类技术,考虑到不同的特征选择方法造成的分类结果的差异,分别在相同和不同的分类样本的条件下测试了4种特征选择方法,研究得出TFIDF方法的突出优点,并总结了合适的特征选择方法对于应用到不同的分类系统的重要性。

Accurate classification for a healthy network environment is of crucial importance.Based on the above background, we choose an ideal effect of the SVM classification technique. Considering the different feature selection methods of the classification results of difference,respectively under the condition of the same and different classification samples tested four feature selection methods,research the prominent importance of TFIDF.And we include that selecting the appropriate feature selection method for application to the different classification system is very important.