Uncovering collusive spammers in Chinese review websites

Xu, Chang; Jie, Zhang; Chang, Kuiyu; Long, Chong

doi:10.1145/2505515.2505700

Cited by 110 publications

(86 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This indicates that the proposed features are superior to these two traditional features. Though the F1 − measure value of the proposed does not outperform that of collusive behavioralbased features, collusive behavioral-based features need to find the group of each user first, which makes it harder to calculate them than our proposed features [32]. All in all, the proposed features are able to detect attackers in the reallife situation, which proves the practical value of our proposed detection method.…”

Section: Practical Value Of the Proposed Detection Methodsmentioning

confidence: 80%

“…The MovieLens100K dataset consists of the ratings from 943 users on 1682 items, with a rating frequency not less than 20 for each item [4]. The Amazon review dataset is crawled from Amazon.cn till August 20, 2012, which contains 1205125 reviews written by 645072 reviewers on 136785 products [32]. Each review has 6 attributes: ReviewerID, ProductID, Product Brand, Rating, Date and Review Text.…”

Section: Experimental Methodsmentioning

confidence: 99%

“…The Amazon review dataset is collected from Amazon China † by Xu etc [32]. This dataset includes 1205125 reviews written by both normal users and attackers.…”

Section: Practical Value Of the Proposed Detection Methodsmentioning

confidence: 99%

“…More detail about these three features can be found in [32]. KNN and SVM are used to take advantage of features to form detection methods.…”

Section: Practical Value Of the Proposed Detection Methodsmentioning

confidence: 99%

See 3 more Smart Citations

Shilling Attack Detection in Recommender Systems via Selecting Patterns Analysis

Gao

Hua

et al. 2016

IEICE Trans. Inf. & Syst.

View full text Add to dashboard Cite

SUMMARYCollaborative filtering (CF) has been widely used in recommender systems to generate personalized recommendations. However, recommender systems using CF are vulnerable to shilling attacks, in which attackers inject fake profiles to manipulate recommendation results. Thus, shilling attacks pose a threat to the credibility of recommender systems. Previous studies mainly derive features from characteristics of item ratings in user profiles to detect attackers, but the methods suffer from low accuracy when attackers adopt new rating patterns. To overcome this drawback, we derive features from properties of item popularity in user profiles, which are determined by users' different selecting patterns. This feature extraction method is based on the prior knowledge that attackers select items to rate with man-made rules while normal users do this according to their inner preferences. Then, machine learning classification approaches are exploited to make use of these features to detect and remove attackers. Experiment results on the MovieLens dataset and Amazon review dataset show that our proposed method improves detection performance. In addition, the results justify the practical value of features derived from selecting patterns.

show abstract

Section: Practical Value Of the Proposed Detection Methodsmentioning

confidence: 80%

Section: Experimental Methodsmentioning

confidence: 99%

“…The Amazon review dataset is collected from Amazon China † by Xu etc [32]. This dataset includes 1205125 reviews written by both normal users and attackers.…”

Section: Practical Value Of the Proposed Detection Methodsmentioning

confidence: 99%

“…More detail about these three features can be found in [32]. KNN and SVM are used to take advantage of features to form detection methods.…”

Section: Practical Value Of the Proposed Detection Methodsmentioning

confidence: 99%

See 2 more Smart Citations

Shilling Attack Detection in Recommender Systems via Selecting Patterns Analysis

Gao

Hua

et al. 2016

IEICE Trans. Inf. & Syst.

View full text Add to dashboard Cite

show abstract

“…Compared to prior works, many challenges arise regarding this problem: (1) (Annotation Difficulty) Considering the fact that CQA spamming is usually a collaborative activity, it is difficult to ascertain which contents are deceptive and which ones are legitimate; (2) (Asymmetric Q&A Attributes) In CQA, the questions and answers are asymmetric with different attributes and linguistic structures, which are different from deceptive product reviews [9,10,23] or promotional microblog posts [13] that can be analyzed uniformly; (3) (Unclear Group Base) Previous works group the spamming activities that review multiple common products in the review platforms [19,28,29] and post common URLs or contents in the microblog environments [4]. However, in CQA, there are not any clear existing connections that can group Q&As, because CWers can generate unlimited distinct questions, and the deceptive answers can respond to any of them.…”

Section: Introductionmentioning

confidence: 99%

Detecting Collusive Spamming Activities in Community Question Answering

Liu

Zhou

et al. 2017

Proceedings of the 26th International Conference on World Wide Web

View full text Add to dashboard Cite

Community Question Answering (CQA) portals provide rich sources of information on a variety of topics. However, the authenticity and quality of questions and answers (Q&As) has proven hard to control. In a troubling direction, the widespread growth of crowdsourcing websites has created a large-scale, potentially difficult-to-detect workforce to manipulate malicious contents in CQA. The crowd workers who join the same crowdsourcing task about promotion campaigns in CQA collusively manipulate deceptive Q&As for promoting a target (product or service). The collusive spamming group can fully control the sentiment of the target. How to utilize the structure and the attributes for detecting manipulated Q&As? How to detect the collusive group and leverage the group information for the detection task?To shed light on these research questions, we propose a unified framework to tackle the challenge of detecting collusive spamming activities of CQA. First, we interpret the questions and answers in CQA as two independent networks. Second, we detect collusive question groups and answer groups from these two networks respectively by measuring the similarity of the contents posted within a short duration. Third, using attributes (individuallevel and group-level) and correlations (user-based and contentbased), we proposed a combined factor graph model to detect deceptive Q&As simultaneously by combining two independent factor graphs. With a large-scale practical data set, we find that the proposed framework can detect deceptive contents at early stage, and outperforms a number of competitive baselines.

show abstract

Spotting review spammer groups: A cosine pattern and network based method

Zhang

Cao

et al. 2018

Concurrency and Computation

View full text Add to dashboard Cite

Nowadays, online product reviews strongly influence the purchase decision of consumers in e-commerce platforms. Driven by the immense financial profits, review spammers deliberately post fake reviews to promote or demote their target products. Some spammers are even organized as groups to work together and try to take total control of the sentiment on their target products. To detect such spammer groups, most previous works exploit frequent itemset mining (FIM) to find spammer group candidates and then use unsupervised spamicity ranking methods to identify real spammer groups. However, these methods usually suffer from the problem of threshold setting, ie, high support value finding fewer groups while low support value leading to more coincidentally generated groups and computational inefficiency. Moreover, the unsupervised methods are not able to make good use of labeled instances which are actually obtainable in practice. In this paper, we propose CONSGD, a cosine pattern and heterogeneous information network-based spammer group detecting method. Specifically, the CONSGD uses cosine pattern mining (CPM) to discover tight spammer group candidates with a respective low support value, where the cosine threshold is utilized to avoid coincidentally generated groups. Moreover, CONSGD employs heterogeneous information network classification to identify the real spammer groups, which could utilize the labeled instances and do not rely to the assumption of independent instances. Experiments on real-life dataset show that our proposed CONSGD is effective and outperforms the state-of-the-art spammer group detection methods. KEYWORDScosine pattern, heterogeneous information network, review spammer group detecting, tight spammer group INTRODUCTIONNowadays, the consumers often read product reviews before they buy the product in e-commerce platforms. When the reviews of one product are positive, the consumer will be very likely to buy it. On the contrary, if most of the reviews are negative, the consumer will discard to buy the product and choose another one. Driven by the immense financial profits, some review spammers (also called opinion spammer) post fake reviews to promote their products or to demote their competitors' products. As the manipulated reviews are misleading more and more consumers in their decision making of purchase, detecting such fake reviews or their posters, ie, spammers, becomes a pressing issue.Ever since the problem of review spam detecting was proposed by Jindal and Liu, 1 many efforts have been done for fake review or individual review spammer detecting. 2-6 Recently, however, many spammers are organized to post fake reviews together to promote or demote some target products. These collusive spammers, also called spammer group, can be highly damaging as they have many accounts to review one product and therefore could take total control of the sentiment on this product. Meanwhile, the spammer group has more manpower to imperceptibly manipulate reviews so that each group member may no longer appear ...

show abstract

Uncovering collusive spammers in Chinese review websites

Cited by 110 publications

References 11 publications

Shilling Attack Detection in Recommender Systems via Selecting Patterns Analysis

Shilling Attack Detection in Recommender Systems via Selecting Patterns Analysis

Detecting Collusive Spamming Activities in Community Question Answering

Spotting review spammer groups: A cosine pattern and network based method

Contact Info

Product

Resources

About