Detect Professional Malicious User with Metric Learning in Recommender Systems

  title={Detect Professional Malicious User with Metric Learning in Recommender Systems},
  author={Yuanbo Xu and Yongjian Yang and En Wang and Fuzhen Zhuang and Hui Xiong},
—In e-commerce, online retailers are usually suffering from professional malicious users (PMUs), who utilize negative reviews and low ratings to their consumed products on purpose to threaten the retailers for illegal profits. PMUs are difficult to be detected because they utilize masking strategies to disguise themselves as normal users. Specifically, there are three challenges for PMU detection: 1) professional malicious users do not conduct any abnormal or illegal interactions (they never… 
1 Citations
Human Origin-Destination Flow Prediction Based on Large Scale Mobile Signal Data
Large-scale mobile phone signal data is used to achieve citywide human OD flow prediction between the coverage of varying signal base stations and a TGCN model combined with a graph fusion module is adopted to pretrain the dynamic population distribution prediction task.


Shilling attack detection utilizing semi-supervised learning method for collaborative recommender system
Semi-SAD, a new semi-supervised learning based shilling attack detection algorithm is proposed to take advantage of only a small number of labeled users in most of the practical recommender systems, while a large number of users are unlabeled because it is expensive to obtain their identities.
HySAD: a semi-supervised hybrid shilling attack detector for trustworthy product recommendation
A Hybrid Shilling Attack Detector, or HySAD for short, is presented, which introduces MC-Relief to select effective detection metrics, and Semi-supervised Naive Bayes to precisely separate Random-Filling model attackers and Average-Filler model attackers from normal users.
Opinion Fraud Detection in Online Reviews by Network Effects
A fast and effective framework for spotting fraudsters and fake reviews in online review datasets, FRAUDEAGLE, which exploits the network effect among reviewers and products and is scalable to large datasets as its run time grows linearly with network size.
Spammers Detection from Product Reviews: A Hybrid Model
Experimental results on movie data sets with shilling injection show that hPSD outperforms several state-of-the-art baseline methods, and shows great potential in detecting hidden spammers as well as their underlying employers from a real-life Amazon data set.
Shilling Attacks Detection in Recommender Systems Based on Target Item Analysis
The experimental results show that the detection model based on target item analysis is an effective approach for detecting shilling attacks and the use of statistical metrics to detect rating patterns of attackers and group characteristics in attack profiles is studied.
Detecting Spammers in E-Commerce Website via Spectrum Features of User Relation Graph
A novel user relation graph model based on a bipartite graph built directly from the review data is proposed and two novel algorithms called Finding Abnormal Dimensions by Kurtosis function (FADK) and FADSW to find small groups of spammers in the large user-relation graph are introduced.
Slanderous user detection with modified recurrent neural networks in recommender system
Detecting shilling attacks in social recommender systems based on time series analysis and trust features
Characterizing and Detecting Malicious Accounts in Privacy-Centric Mobile Social Networks: A Case Study
This work proposes dozens of new features and leverage machine learning to detect malicious accounts in a new type of OSN, called privacy-centric mobile social network (PC-MSN), such as KakaoTalk and LINE, which has attracted billions of users recently.