摘要
微博中大量的垃圾互粉现象给影响力甄别和用户关系强度分析带来了挑战。提出一种基于多特征的垃圾互粉快速检测方法,通过提取用户个人信息、用户间关系和用户行为等多种类别的特征,训练了一个垃圾互粉用户和正常用户的分类系统。在新浪微博上的实验结果显示,该方法检测垃圾互粉用户的有效性达到80%以上。
Spam mutual concems in micro-blogs bring considerable challenges to the determination of influence of bloggers and user relation strength. A method to detect the spare mutual concerns is presented. Based on features extracted from user infomaafion, relation among user and user activity, a system is Irained to distinguish spammers from normal users. Experiments on real data fi＇om Sina Weibo （Micro-blog） show that the method is effective with a distinguishing rate of above 80%.
出处
《中国科技论文》
CAS
北大核心
2012年第7期548-551,共4页
China Sciencepaper
关键词
微博
新浪微博
垃圾互粉
作弊检测
micro-blog
SinaWeibo
spammutualconcem
spare detection
