摘要： 为更有效地甄别微博僵尸粉，提出一种基于微博注册用户名特征提取的智能分类方法。以新浪微博作为研究平台，通过对微博用户数据进行分析，构建标准匹配库，提取用户名特征向量，再分别利用支持向量机(SVM)和人工神经网络 (ANN)方法对特征集合进行分类。实验结果表明，将用户名特征提取与SVM、ANN相结合，僵尸粉甄别准确率均高于92%。

关键词: 微博, 僵尸粉, 特征提取, 智能分类, 支持向量机, 人工神经网络

Abstract: In order to recognize zombie fans of micro-blog, this paper proposes a new intelligent classification method based on the characteristics of registered usernames. Taking Sina micro-blog as research platform, by analyzing its giant user database, a standard matching library is constructed, and the user characteristics set is extracted to be trained and predicted with Supported Vector Machine(SVM) and Artificial Neural Network(ANN). Experimental results show that combing user characteristics, SVM and ANN can effectively solve the problem of classification of usernames, and the accuracy ratio of recognition is higher than 92%.

Key words: micro-blog, zombie fans, feature extraction, intelligent classification, Support Vector Machine(SVM), Artificial Neural Network(ANN)

中图分类号: 

• TP391