{"paper_id":"2312.13896","title":"Comparative Evaluation of Anomaly Detection Methods for Fraud Detection in Online Credit Card Payments","authors":["Hugo Thimonier","Fabrice Popineau","Arpad Rimmel","Bich-Liên Doan","Fabrice Daniel"],"paper_date":"2023-12-21","paper_type":"empirical","paper_url":"https://arxiv.org/pdf/2312.13896v1","page_url":"https://thequant.space/flowcharts/comparative-evaluation-of-anomaly-detection-methods-for-frau/","scores":{"empirical_rigor":8.5,"math_complexity":5,"quadrant":"Holy Grail"},"proposes":{"status":"available","statement":"The paper empirically evaluates recent anomaly detection methods against LightGBM for fraud detection on real online credit card payment data.","quote":"Our findings reveal that LightGBM exhibits significantly superior performance across all evaluated metrics but suffers more from distribution shifts than AD methods.","summary":"This study explores the application of anomaly detection (AD) methods in imbalanced learning tasks, focusing on fraud detection using real online credit card payment data. We assess the performance of several recent AD methods and compare their effectiveness against standard supervised learning…","strategy":null},"assumptions":{"status":"available","items":[{"assumption":"Only fraudulent behaviors change over time, so AD models should be more robust to distribution shift than supervised approaches.","type":"model","stated":"explicit","quote":"assuming only fraudulent behaviors change over time, AD models should be more robust to distribution shift than standard supervised approaches."},{"assumption":"If the normal distribution is well characterized, AD models can always exclude new types of anomalies.","type":"model","stated":"explicit","quote":"if the normal distribution is well characterized, they should always be able to exclude new types of anomalies."}]},"datasets":{"status":"available","items":[{"name":"Online credit card payments dataset","provider":"major French bank","kind":"other","access":"proprietary","period":"2018-01 to 2021-12","universe":"online credit card transactions","quote":"We dispose of a labeled dataset of online credit card payments made available by a major French bank. Our dataset contains 145 features"}],"known":[],"sample":{"start":null,"end":null,"market":null,"asset_class":null,"frequency":null},"data_resolution":"unclear"},"transaction_costs":{"status":"not_applicable","modeled":"n/a","headline_net_of_costs":null,"detail":null,"quote":null,"source":"front_matter","rule":"R3","candidates":{"front_matter":"n/a","evidence_card":"n/a","evidence_card_quoted":false},"meaning":"the paper reports no trading or execution result"},"results":{"status":"available","evaluation_type":"forecast-accuracy","findings":[{"finding":"LightGBM consistently outperforms all tested AD methods across various evaluation metrics.","kind":"performance","quote":"Our findings reveal that LightGBM exhibits significantly superior performance across all evaluated metrics but suffers more from distribution shifts than AD methods."},{"finding":"LightGBM suffers more performance degradation due to distribution shift than AD methods.","kind":"negative","quote":"show that GBDT suffer significantly more from distribution shift than AD methods while displaying substantially better fraud detection performance than all tested AD methods."},{"finding":"LightGBM captures the majority of frauds detected by AD methods, limiting the benefit of ensembling.","kind":"performance","quote":"LightGBM also captures the majority of frauds detected by AD methods. This observation challenges the potential benefits of ensemble methods to combine supervised, and AD approaches to enhance performance."},{"finding":"ECOD detected 20.41% of fraud cases not flagged by LightGBM, but 89.02% of ECOD‑detected frauds were also detected by LightGBM.","kind":"performance","quote":"that ECOD did not flag. Notably, the 3.28% represents 10.98% of the fraud cases detected by ECOD. In other words, 89.02% of the fraud cases detected by ECOD were also detected by LightGBM."}],"reported_performance":null,"out_of_sample":{"method":"holdout","headline_out_of_sample":null}},"independent_evidence":{"status":"none_found","supports":[],"contradicts":[],"qualifies":[],"compared_with":33,"citations":{"cited_by_in_corpus":0,"by_independent_authors":0,"how_cited":{"used_as_baseline":0,"uses_method_or_data":0,"extends":0,"discusses_findings":0,"mentions":0,"unclassified":0},"independent_papers":[],"scope":"Citations between papers in this corpus only; citations from outside it are not counted."}},"implementation":{"status":"available","public":false,"code_url":null,"note":"No repository presented by the authors as their own was found in the paper. Absence here is not proof that no code exists."},"replication":{"status":"none_found","card":null,"note":"No replication of this paper has been run by thequant.space yet."},"extraction":{"text_source":"fulltext","model":"openai/gpt-oss-120b","version":1,"extracted_at":"2026-11-14T18:17:22Z"},"provenance":{"claims":"Everything here is what the authors state in the paper. Nothing was re-run or reproduced by the platform.","quotes":"Assumptions, datasets, cost detail and findings were extracted by a language model. Every quote is verbatim paper text. Where an item has `support`, \"full\" means every figure in it appears in the quoted sentences and \"partial\" means a figure is stated elsewhere in the paper; items with a figure found nowhere in the paper were discarded.","datasets_known":"datasets.known lists well-known vendors and public datasets by a stable id, found by a dictionary over the full text where the sentence says the data is used; it is independent of the model and covers only names in the dictionary.","flags":"transaction_costs.modeled, results.evaluation_type and results.out_of_sample.method are platform classifications (structured-decision models); see https://thequant.space/score-guide/#penalty-flags","comparisons":"supports / contradicts / qualifies are a language model's reading of two papers' quoted findings, restricted to closely related papers that share no author. They are pointers to read, not verified replications, and coverage is partial.","implementation":"code_url is a repository the paper itself presents as its own (rule-scored, ambiguous cases judged by a model), checked for liveness when it was recorded."},"docs":"https://thequant.space/agents/"}