协同过滤算法的欺诈攻击分析

Analysis of Shilling Attacks on Collaborative Filtering Algorithms

协同过滤是个性化推荐系统最常用的一种技术,被广泛应用于电子商务,但它对用户概貌信息较为敏感,欺诈攻击者很容易通过注入有偏差的用户概貌信息,人为干预推荐系统的结果。针对这个问题,实验分析基于奇异值分解(SVD)的协同过滤算法在随机攻击模型下的性能表现,并以三种评估指标分析不同攻击规模和填充规模对攻击效率的影响。

Collaborative filtering is being a major tool of the personalized recommender systems and widely used in e-commence,but it is so sensitive to user profiles,that shilling attackers can easily inject biased profiles in an attempt to intervene the result of the recommender systems artificially.This paper analyzes the attack effectiveness of random attack model on a SVD-based collaborative filtering algorithm,and the performances of attack models with different attack sizes and fill sizes using three evaluation parameters.