针对现有托攻击检测算法在攻击强度较小时正确识别率低、实施成本高的缺点,在基于主成分分析变量选择(Variableselection using principal component analysis,PCA Var Select)的基础上,提出一种无需事先知道攻击强度的攻击检测算法.首先,利用PCA Var Select方法得到每个用户的主成分值;然后,选取主成分值较小范围的用户集,用来确定被攻击物品和评分向量长度;最后,将对被攻击物品评高分的用户集、具有嫌疑评分向量长度的用户集、主成分值较大范围的用户集三者取交,得到攻击用户集.经过实验验证,该算法在去除PCA Var Select对先验知识依赖的同时,同时也提高了准确率,在面对小规模攻击强度时表现依然良好.
Purpose:The paper aims to build an index model for measuring microblog users’influence by taking microbloggers of Sina Weibo as a research sample.Design/methodology/approach:Our user influence index model emphasizes link analysis and user activities in the microblogging network.We conduct experiments to investigate the performance of our model by using data crawled from Sina Weibo.Findings:User influence is correlated to the attention that a user has received from his/her audience,the user’s activities and his/her tweets’influence.Experimental results show that our model can reflect microbloggers’influence in a more reasonable way.Research limitations:More factors need to be considered to identify different influential users at different time periods.Practical implications:The results of the study provide us with insights both into the way to measure microblog users’influence and to rank users based on their influence.Originality/value:By combining link analysis and user activities,this index model can reduce the impact of dummy follower accounts on user influence,reflecting a user’s real influence in the microblog system.
