中文说明：
针对SVM法线特征筛选算法仅考虑法线对特征筛选的贡献,而忽略了特征分布对特征筛选的贡献的不足,在对SVM法线算法进行分析的基础上,基于特征在正、负例中出现概率的不同提出了加权SVM法线算法,该算法考虑到了法线和特征的分布.通过试验可以看出,在使用较小的特征空间时,与SVM法线算法和信息增益算法相比,加权SVM法线算法具有更好的特征筛选性能。
English Description：
Aiming at the deficiency that SVM normal feature screening algorithm only considers the contribution of normal to feature screening and ignores the contribution of feature distribution to feature screening, based on the analysis of SVM normal algorithm and the difference of feature occurrence probability in positive and negative cases, a weighted SVM normal algorithm is proposed, which takes into account the distribution of normal and feature Experiments show that when using a small feature space, weighted SVM normal algorithm has better feature screening performance than SVM normal algorithm and information gain algorithm