{"url":"/dataset/yelpchi","name":"Yelp-Fraud","full_name":"Multi-relational Graph Dataset for Yelp Spam Review Detection","description_markdown":"Yelp-Fraud is a multi-relational graph dataset built upon the [Yelp spam review dataset](http://odds.cs.stonybrook.edu/yelpchi-dataset/), which can be used in evaluating graph-based node classification, fraud detection, and anomaly detection models.\r\n\r\n- **Dataset Statistics**\r\n\r\n| # Nodes  |  %Fraud Nodes (Class=1) |\r\n|-------|--------|\r\n|  45,954 | 14.5   | \r\n\r\n| Relation  | # Edges |\r\n|--------|--------|\r\n  |   R-U-R    |  49,315 |\r\n |  R-T-R  |  573,616  |\r\n|  R-S-R  |  3,402,743 |\r\n |  All |  3,846,979  |\r\n\r\n- **Graph Construction**\r\n\r\nThe Yelp spam review dataset includes hotel and restaurant reviews filtered (spam) and recommended (legitimate) by Yelp. We conduct a spam review detection task on the Yelp-Fraud dataset which is a binary classification task. We take 32 handcrafted features from [SpEagle](http://shebuti.com/wp-content/uploads/2016/06/15-kdd-collectiveopinionspam.pdf) paper as the raw node features for Yelp-Fraud. Based on previous studies which show that opinion fraudsters have connections in user, product, review text, and time, we take reviews as nodes in the graph and design three relations: **1) R-U-R:** it connects reviews posted by the same user; **2) R-S-R:** it connects reviews under the same product with the same star rating (1-5 stars); **3) R-T-R:** it connects two reviews under the same product posted in the same month. \r\n\r\nTo download the dataset, please visit [this](https://github.com/YingtongDou/CARE-GNN) Github repo. For any other questions, please email ytongdou(AT)gmail.com for inquiry.","description_withheld":null,"homepage":"https://github.com/YingtongDou/CARE-GNN","introduced_date":"2020-08-19","introduced_date_note":null,"introduced_by":{"paper":"/paper/enhancing-graph-neural-network-based-fraud","title":"Enhancing Graph Neural Network-based Fraud Detectors against Camouflaged Fraudsters","first_author":"Yingtong Dou","url":null},"license":{"name":"Apache-2.0","url":"https://github.com/YingtongDou/CARE-GNN/blob/master/LICENSE"},"modalities":[{"name":"Graphs","url":"/datasets/modality/graphs"}],"tasks":[{"name":"Anomaly Detection","url":"/task/anomaly-detection","datasets_with_task":"/datasets/task/anomaly-detection"},{"name":"Node Classification","url":"/task/node-classification","datasets_with_task":"/datasets/task/node-classification"},{"name":"Fraud Detection","url":"/task/fraud-detection","datasets_with_task":"/datasets/task/fraud-detection"},{"name":"Graph Anomaly Detection","url":"/task/graph-anomaly-detection","datasets_with_task":"/datasets/task/graph-anomaly-detection"},{"name":"Graph Mining","url":"/task/graph-mining","datasets_with_task":"/datasets/task/graph-mining"}],"languages":[{"name":"English","url":"/datasets/language/english"}],"variants":["Yelp-Fraud"],"data_loaders":[{"repo":"https://github.com/dmlc/dgl","url":"https://docs.dgl.ai/api/python/dgl.data.html#fraud-dataset","frameworks":["pytorch"]},{"repo":"https://github.com/YingtongDou/CARE-GNN","url":"https://github.com/YingtongDou/CARE-GNN","frameworks":["pytorch"]}],"num_papers_in_archive":13,"source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28"},"benchmarks":[{"leaderboard":"/sota/fraud-detection-on-yelp-fraud","task":"Fraud Detection","dataset_variant":"Yelp-Fraud","rows":10,"metrics":["AUC-ROC","Averaged Precision","F1 Macro","G-mean"],"first_row_in_archive_order":{"model":"LEX-GNN","paper":"/paper/lex-gnn-label-exploring-graph-neural-network","metrics":{"AUC-ROC":"96.40","Averaged Precision":"83.56","F1 Macro":"86.35","G-mean":"84.91"},"code_links":[{"title":"wdhyun/LEX-GNN","url":"https://github.com/wdhyun/LEX-GNN"}]},"note":"rows are the archive's own order at snapshot; nothing here re-ranks them"},{"leaderboard":"/sota/node-classification-on-yelpchi","task":"Node Classification","dataset_variant":"Yelp-Fraud","rows":9,"metrics":["AUC-ROC"],"first_row_in_archive_order":{"model":"LEX-GNN","paper":"/paper/lex-gnn-label-exploring-graph-neural-network","metrics":{"AUC-ROC":"96.40"},"code_links":[{"title":"wdhyun/LEX-GNN","url":"https://github.com/wdhyun/LEX-GNN"}]},"note":"rows are the archive's own order at snapshot; nothing here re-ranks them"},{"leaderboard":"/sota/graph-anomaly-detection-on-yelp-fraud","task":"Graph Anomaly Detection","dataset_variant":"Yelp-Fraud","rows":1,"metrics":["AUC"],"first_row_in_archive_order":{"model":"Grad","paper":"/paper/grad-guided-relation-diffusion-generation-for","metrics":{"AUC":"99.08"},"code_links":[{"title":"ai4risk/antifraud","url":"https://github.com/ai4risk/antifraud"},{"title":"Muyiiiii/WWW25-Grad","url":"https://github.com/Muyiiiii/WWW25-Grad"},{"title":"Muyiiiii-HF/WWW25-Grad","url":"https://huggingface.co/Muyiiiii-HF/WWW25-Grad"}]},"note":"rows are the archive's own order at snapshot; nothing here re-ranks them"}],"papers_with_a_benchmark_row":[{"paper":"/paper/grad-guided-relation-diffusion-generation-for","title":"Grad: Guided Relation Diffusion Generation for Graph Augmentation in Graph Fraud Detection","date":"2025-04-22","rows_on_this_dataset":1,"code_links":3,"syntology":null},{"paper":"/paper/semi-supervised-credit-card-fraud-detection-1","title":"Semi-supervised Credit Card Fraud Detection via Attribute-Driven Graph Representation","date":"2024-12-24","rows_on_this_dataset":2,"code_links":2,"syntology":null},{"paper":"/paper/financial-fraud-detection-using-jump","title":"Financial Fraud Detection using Jump-Attentive Graph Neural Networks","date":"2024-11-07","rows_on_this_dataset":1,"code_links":1,"syntology":null},{"paper":"/paper/lex-gnn-label-exploring-graph-neural-network","title":"LEX-GNN: Label-Exploring Graph Neural Network for Accurate Fraud Detection","date":"2024-10-21","rows_on_this_dataset":2,"code_links":1,"syntology":null},{"paper":"/paper/splitgnn-spectral-graph-neural-network-for","title":"SplitGNN: Spectral Graph Neural Network for Fraud Detection against Heterophily","date":"2023-10-21","rows_on_this_dataset":2,"code_links":2,"syntology":null},{"paper":"/paper/bolt-an-automated-deep-learning-framework-for","title":"BOLT: An Automated Deep Learning Framework for Training and Deploying Large-Scale Search and Recommendation Models on Commodity CPU Hardware","date":"2023-03-30","rows_on_this_dataset":2,"code_links":2,"syntology":null},{"paper":"/paper/rlc-gnn-an-improved-deep-architecture-for","title":"RLC-GNN: An Improved Deep Architecture for Spatial-Based Graph Neural Network with Application to Fraud Detection","date":"2021-06-18","rows_on_this_dataset":2,"code_links":0,"syntology":null},{"paper":"/paper/pick-and-choose-a-gnn-based-imbalanced","title":"Pick and Choose: A GNN-based Imbalanced Learning Approach for Fraud Detection","date":"2021-04-19","rows_on_this_dataset":2,"code_links":1,"syntology":null},{"paper":"/paper/reinforced-neighborhood-selection-guided","title":"Reinforced Neighborhood Selection Guided Multi-Relational Graph Neural Networks","date":"2021-04-16","rows_on_this_dataset":2,"code_links":1,"syntology":null},{"paper":"/paper/new-benchmarks-for-learning-on-non","title":"New Benchmarks for Learning on Non-Homophilous Graphs","date":"2021-04-03","rows_on_this_dataset":2,"code_links":1,"syntology":null},{"paper":"/paper/enhancing-graph-neural-network-based-fraud","title":"Enhancing Graph Neural Network-based Fraud Detectors against Camouflaged Fraudsters","date":"2020-08-19","rows_on_this_dataset":2,"code_links":6,"syntology":{"read_at":"2026-09-24T18:15:14+00:00","samples_harvested":17,"samples_ran":1,"samples_unverified":16,"pointer_only_for_licence":0,"claim":"Per-sample execution on synthesized fixtures; not a correctness claim."}}],"syntology_totals":{"read_at":"2026-09-24T18:15:14+00:00","papers_with_samples":1,"samples_harvested":17,"samples_ran":1,"samples_unverified":16,"pointer_only_for_licence":0,"papers_with_no_sample_that_ran":0,"note":"the per-paper counts above, summed; not a rate"},"papers_note":"The archive never published its papers-using-dataset list; these are papers with a leaderboard row on this dataset's benchmarks."}