The Influence of Shilling Attacks with Different Attack Cycles
Feng Jiang, Renli Tian · 2017
Due to the rapid development of the Internet and the explosive growth of information resources, the personalized recommendation technology has attracted extensive attentions as an effective tool to solve the information overload. However, the openness of recommendation systems makes the systems suffer from shilling attackers who achieve the purpose of the attack by injecting false ratings into the systems. Researchers on shilling attacks have proposed several attack models with a variety of possible ratings, and analyzed their costs and attack effects. Injection time is an important factor of the attack, but the current research on the shilling attack has yet taken the factor into consideration. Therefore, this paper proposes eight shilling attack models based on the different strategies of the injection time of shilling attacks, and analyzes the cost and effect of each attack model under different attack cycles. Experiments show whether equal time attacks or random time attacks are significant influenced by attack cycles.