Micro-blog spammer detection based on characteristics of social behaviors
Jianbo Wang, Hua Li, Jianping Zhao · 2017
Detection of spammer is being paid a great attention in the area of data mining and social media. In this paper, we propose a method of detecting micro-blog spammers based on characteristics of social behaviors. We get data through microblog API and crawler. We turn the research perspective from its media meaning to its social meaning. According to characteristics of social behaviors, this paper classifies the spammers into 2 categories: zombie fans and water army. We use 3 machine learning methods: Support Vector Machine (SVM), Decision Trees (DT) and Naive Bayes (NB) to detect spammers. The experimental results show that we achieved accuracy of 96.19% of legitimate users, 98.45% of zombie fans and 93.62% of water army based on the micro-blog data sets.