Review spam detection via temporal pattern discovery

Online reviews play a crucial role in today's electronic commerce. It is desirable for a customer to read reviews of products or stores before making the decision of what or from where to buy. Due to the pervasive spam reviews, customers can be misled to buy low-quality products, while decent stores can be defamed by malicious reviews. We observe that, in reality, a great portion (> 90% in the data we study) of the reviewers write only one review (singleton review). These reviews are so enormous in number that they can almost determine a store's rating and impression. However, existing methods did not examine this larger part of the reviews. Are most of these singleton reviews truthful ones? If not, how to detect spam reviews in singleton reviews? We call this problem singleton review spam detection. To address this problem, we observe that the normal reviewers' arrival pattern is stable and uncorrelated to their rating pattern temporally. In contrast, spam attacks are usually bursty and either positively or negatively correlated to the rating. Thus, we propose to detect such attacks via unusually correlated temporal patterns. We identify and construct multidimensional time series based on aggregate statistics, in order to depict and mine such correlations. In this way, the singleton review spam detection problem is mapped to a abnormally correlated pattern detection problem. We propose a hierarchical algorithm to robustly detect the time windows where such attacks are likely to have happened. The algorithm also pinpoints such windows in different time resolutions to facilitate faster human inspection. Experimental results show that the proposed method is effective in detecting singleton review attacks. We discover that singleton review is a significant source of spam reviews and largely affects the ratings of online stores.