Quantifying Assimilate-Contrast Effects in Online Rating Systems

The unbiasedness of online product ratings, an important property to ensure that users’ ratings indeed reflect their true evaluations to products, is vital both in shaping consumer purchase decisions and providing reliable recommendations. Recent experimental studies showed that distortions from historical ratings would ruin the unbiasedness of subsequent ratings. How to “discover” the distortions from historical ratings in each single rating (or at the micro-level), and perform the “debiasing operations” in real rating systems are the main objectives of this work. Using 42 million real customer ratings, we first show that users either “assimilate” or “contrast” to historical ratings under different scenarios: users conform to historical ratings if historical ratings are not far from the product quality (assimilation), while users deviate from historical ratings if historical ratings are significantly different from the product quality (contrast). This phenomenon can be explained by the well-known psychological argument: the “Assimilate-Contrast” theory. However, none of the existing works on modeling historical ratings’ influence have taken this into account, and this motivates us to propose the Histori- cal Influence Aware Latent Factor Model (HIALF), the first model for real rating systems to capture and mitigate historical distortions in each single rating. HIALF also allows us to study the influence patterns of historical ratings from a modeling perspective, and it perfectly matches the assimilation and contrast effects we previously observed. Also, HIALF achieves significant improvements in predicting subsequent ratings, and accurately predicts the relationships revealed in previous empirical measurements on real ratings. Finally, we show that HIALF can contribute to better recommendations by decoupling users’ real preference from distorted ratings, and reveal the intrinsic product quality for wiser consumer purchase decisions.

Download Full-text

Detecting Anomalous Ratings in Collaborative Filtering Recommender Systems

International Journal of Digital Crime and Forensics ◽

10.4018/ijdcf.2016040102 ◽

2016 ◽

Vol 8 (2) ◽

pp. 16-26 ◽

Cited By ~ 1

Author(s):

Zhihai Yang ◽

Zhongmin Cai

Keyword(s):

Collaborative Filtering ◽

Recommender Systems ◽

Graph Mining ◽

Scale Up ◽

Detection Methods ◽

Detection Accuracy ◽

Rating Systems ◽

Rating Data ◽

Online Rating ◽

Two Stages

Online rating data is ubiquitous on existing popular E-commerce websites such as Amazon, Yelp etc., which influences deeply the following customer choices about products used by E-businessman. Collaborative filtering recommender systems (CFRSs) play crucial role in rating systems. Since CFRSs are highly vulnerable to “shilling” attacks, it is common occurrence that attackers contaminate the rating systems with malicious rates to achieve their attack intentions. Despite detection methods based on such attacks have received much attention, the problem of detection accuracy remains largely unsolved. Moreover, few can scale up to handle large networks. This paper proposes a fast and effective detection method which combines two stages to find out abnormal users. Firstly, the manuscript employs a graph mining method to spot automatically suspicious nodes in a constructed graph with millions of nodes. And then, this manuscript continue to determine abnormal users by exploiting suspected target items based on the result of first stage. Experiments evaluate the effectiveness of the method.

Download Full-text