Comparative Study of Data-Level Imbalance Handling Techniques with Ensemble Models for Credit Card Fraud Detection

Comparative Study of Data-Level Imbalance Handling Techniques with Ensemble Models for Credit Card Fraud Detection | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Method Article Comparative Study of Data-Level Imbalance Handling Techniques with Ensemble Models for Credit Card Fraud Detection Noha Youssef, Malak Gaballa, Masa Tantawy, Moustafa El Mahdy This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7004067/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Credit card fraud detection remains a challenging problem due to extreme class imbalance, where fraudulent transactions constitute only a small fraction of total activity. This study presents a comparative analysis of six widely used data-level imbalance handling techniques; Random Undersampling, TomekLinks, Random Oversampling, SMOTE, SMOTETomek, and SMOTEENN paired with three ensemble classifiers: Random Forest, XGBoost, and LightGBM. Experiments were conducted on a large, publicly available synthetic transaction dataset using both 5-fold cross-validation and an 80/20 train-test split. Performance was assessed using AUC-ROC, F1-score, precision, and recall. XGBoost with TomekLinks achieved the highest F1-score of 0.8763 and an AUC of 0.9986, outperforming all other combinations. A paired t-test comparing F1-scores across sampling methods showed that XGBoost significantly outperforms LightGBM ( \(p\) = 0.013), with a mean F1-score difference of 0.1502. These results highlight XGBoost as a strong candidate for fraud detection under class imbalance, particularly when paired with TomekLinks resampling. The findings underscore the importance of selecting appropriate resampling strategies and suggest that model choice can be informed by both statistical evidence and practical deployment considerations. Credit Card Fraud Detection Data-Level Imbalance Handling Machine Learning Models Class Imbalance Solutions Fraud Prevention Techniques Random Oversampling SMOTE Financial Security Digital Payment Systems Algorithmic Accuracy Full Text Additional Declarations No competing interests reported. Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-7004067","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Method Article","associatedPublications":[],"authors":[{"id":499227552,"identity":"233adc40-789a-4602-abfa-8612012e2db8","order_by":0,"name":"Noha Youssef","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA9klEQVRIiWNgGAWjYDAD9vYGJB4PMVp4zhyAstiI1nIjgUgt5uxnDB/z/KmT45F8/EziR0Vt4nb5BsYHb9twa7HsyTE25m07bMwjnWYm2XPmeOLONgZmw7l4tBgcyDGT5m04kLhfOsFMmrHtWOKGYwxs0rz4tJx/YyYNdFh9j+Txb9KM/8Ba2H/j1XIDaAsPG3MCjwQP0JaGGrAtzPi1PCsGuvywYQ9PTrFlz7EDxjvbEpsl55zD57DkjQ/e/KmT52E/vvHGj5o62e3Mhw9+eFOGWwsDA4cBMu+w4wYGxgZ86oGA/QEyr87eAKuqUTAKRsEoGMkAAI+6UsSdjU09AAAAAElFTkSuQmCC","orcid":"","institution":"The American University in Cairo","correspondingAuthor":true,"prefix":"","firstName":"Noha","middleName":"","lastName":"Youssef","suffix":""},{"id":499227553,"identity":"04b479e0-f3a6-4db4-a631-a3002a4a2f1e","order_by":1,"name":"Malak Gaballa","email":"","orcid":"","institution":"The American University in Cairo","correspondingAuthor":false,"prefix":"","firstName":"Malak","middleName":"","lastName":"Gaballa","suffix":""},{"id":499227554,"identity":"020ac556-de92-4be0-9d99-3695bcafca8c","order_by":2,"name":"Masa Tantawy","email":"","orcid":"","institution":"The American University in Cairo","correspondingAuthor":false,"prefix":"","firstName":"Masa","middleName":"","lastName":"Tantawy","suffix":""},{"id":499227555,"identity":"d4116ca9-07d4-4a16-84a6-c1132ebc2514","order_by":3,"name":"Moustafa El Mahdy","email":"","orcid":"","institution":"The American University in Cairo","correspondingAuthor":false,"prefix":"","firstName":"Moustafa","middleName":"El","lastName":"Mahdy","suffix":""}],"badges":[],"createdAt":"2025-06-29 17:08:13","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7004067/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7004067/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":93607936,"identity":"ae731383-59b5-41d7-85e8-9d657f6872e1","added_by":"auto","created_at":"2025-10-15 15:47:18","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1004574,"visible":true,"origin":"","legend":"","description":"","filename":"FraudDetection.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7004067/v1_covered_d664ab2d-c539-4aaf-8eef-8cb534a2e834.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Comparative Study of Data-Level Imbalance Handling Techniques with Ensemble Models for Credit Card Fraud Detection","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":true,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":true,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Credit Card Fraud Detection, Data-Level Imbalance Handling, Machine Learning Models, Class Imbalance Solutions, Fraud Prevention Techniques, Random Oversampling, SMOTE, Financial Security, Digital Payment Systems, Algorithmic Accuracy","lastPublishedDoi":"10.21203/rs.3.rs-7004067/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7004067/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eCredit card fraud detection remains a challenging problem due to extreme class imbalance, where fraudulent transactions constitute only a small fraction of total activity. This study presents a comparative analysis of six widely used data-level imbalance handling techniques; Random Undersampling, TomekLinks, Random Oversampling, SMOTE, SMOTETomek, and SMOTEENN paired with three ensemble classifiers: Random Forest, XGBoost, and LightGBM. Experiments were conducted on a large, publicly available synthetic transaction dataset using both 5-fold cross-validation and an 80/20 train-test split. Performance was assessed using AUC-ROC, F1-score, precision, and recall. XGBoost with TomekLinks achieved the highest F1-score of 0.8763 and an AUC of 0.9986, outperforming all other combinations. A paired t-test comparing F1-scores across sampling methods showed that XGBoost significantly outperforms LightGBM (\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(p\\)\u003c/span\u003e\u003c/span\u003e = 0.013), with a mean F1-score difference of 0.1502. These results highlight XGBoost as a strong candidate for fraud detection under class imbalance, particularly when paired with TomekLinks resampling. The findings underscore the importance of selecting appropriate resampling strategies and suggest that model choice can be informed by both statistical evidence and practical deployment considerations.\u003c/p\u003e","manuscriptTitle":"Comparative Study of Data-Level Imbalance Handling Techniques with Ensemble Models for Credit Card Fraud Detection","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-08-14 04:18:51","doi":"10.21203/rs.3.rs-7004067/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"0f7c8e0c-6e45-4565-9d35-387829f217a6","owner":[],"postedDate":"August 14th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2025-10-15T15:23:10+00:00","versionOfRecord":[],"versionCreatedAt":"2025-08-14 04:18:51","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7004067","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7004067","identity":"rs-7004067","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}