基于复合分类模型的社交网络恶意用户识别方法

MALICIOUS USERS IDENTIFICATION IN SOCIAL NETWORK BASED ON COMPOSITE CLASSIFICATION MODEL

社交网络近年发展迅速,微博类社交网络的用户数目及规模急剧增大的同时也带来了诸多安全问题,为了保护用户的隐私和个人、集体的利益,需要针对这些恶意行为进行识别并对恶意用户进行处理。提出一种采用复合分类模型对用户进行分类的方法,并开发了一个对微博类社交网络用户进行分类的系统。通过研究用户的属性和行为特点,比较属性间的相关性,从两方面兼顾了分类的准确性和效率。

While having sharp increase in users and network size as in social network of microblogging, the rapid development of social network in recent years also brings lots of security problems. To protect user privacy, personal and collective interest against violations of these security issues, it is necessary to identify malicious behaviours and deal with malicious users. This paper presents a new method for classifying social network users on composite classification model and develops a system to classify users in social network of microblogging. The system analyses many features of the properties and behaviours of users and compares the correlation between the properties, and is able to take the account of both accuracy and efficiency.