Abstract | ||
---|---|---|
Online product reviews are becoming the second most trusted source of product information, second only to recommendations from family and friends, because consumers think that online product reviews reflect recommendations of "real" people. However, in order to maximize the impact, some merchants organize a group of fraudulent reviewers to post a lot of fraudulent reviews that mislead consumers, which is called review spammer group. Solutions for review spammer group detection are very limited, and most solutions focus on static review networks. In this paper, we propose an online two-step framework, called OGSpam, detecting review spammer groups in dynamic review networks. By model a dynamic review network as an initial static review network with an infinite change review stream, our framework first detects reviewer groups on the initial static review network (first snapshot) based on classical Clique Percolation Method (CPM). Then, it detects reviewer groups on snapshot T+1 using reviewer network at T+1 and reviewer groups at T. The experimental results on two real-world review datasets illustrate the efficiency and effectiveness of our framework. To the best of our knowledge, this is the first time to detect review spammer group in dynamic review network.
|
Year | DOI | Venue |
---|---|---|
2019 | 10.1145/3321408.3323077 | Proceedings of the ACM Turing Celebration Conference - China |
Keywords | Field | DocType |
clique percolation method, dynamic review network, online learning, review spam, spammer group detection | Computer science,Artificial intelligence,Machine learning,Spamming | Conference |
ISBN | Citations | PageRank |
978-1-4503-7158-2 | 0 | 0.34 |
References | Authors | |
0 | 4 |
Name | Order | Citations | PageRank |
---|---|---|---|
Mengxiao Hu | 1 | 0 | 1.69 |
Guangxia Xu | 2 | 42 | 9.46 |
Chuang Ma | 3 | 16 | 7.00 |
Mahmoud Daneshmand | 4 | 345 | 46.70 |