基于账户信用评价的恶意发帖检测系统研究

摘要

At present,the malicious posting detection system deals with the post itself,but ignores that the post is the user′s subjective behavior. Proceeding from the research of the user′s subjective behavior,an account credit model was estab⁃lished and a malicious post detection system based on account credit evaluation was designed to eliminate the phenomenon. The main influencing factors of account credit model are analyzed and discretized for account credit quantification. By establishing the model account credit,the accounts' credit standing is evaluated and classified to effectively discover malicious accounts,pre⁃dict posting behavior,and realize strict monitoring of malicious accounts. The malicious posting detection system is dynamically adjusted according to the credit feedback of the account. The validity of the malicious post detection system was verified through the experiment.%恶意发帖检测系统处理的主体是帖子，往往忽略发帖是用户的主观行为。针对这一现象，从研究用户的主观行为出发，建立账户信用模型，设计基于账户信用评价的恶意发帖检测系统。分析账户信用模型的主要影响因素，将其离散化，量化账户的信用。通过建立账户信用模型，对账户进行信用评价和分类，有效发现恶意账户，预测发帖行为，对恶意账号进行严格监控，并根据账户信用影响反馈，动态调整恶意发帖检测系统。通过实验，验证了恶意发帖检测系统的有效性。