摘要
利用改进的支持向量机进行垃圾邮件识别,先对样本进行SVD降维,再结合有标记与无标记样本进行直推式支持向量机训练,实验证明获得的分类器具有较好识别效果.
In this paper,the author uses the improved support vector machines for email classification.Firstly,use SVD dimensionality reduction of the samples,then combined with labeled and unlabeled samples for transductive support vector machine training.Experiments show that the classification results have better identify spam.
出处
《韶关学院学报》
2012年第2期13-16,共4页
Journal of Shaoguan University
