[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-126580-en":3,"doc-seo-126580-105":31,"detail-sidebar-cat-0-en-105":92},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":20,"is_deleted":4,"is_public":21,"is_downloadable":21,"audit_status":21,"page_count":22,"language":23,"language_code":24,"site_id":25,"html_lang":24,"table_of_contents":26,"faqs":27,"seo_title":28,"seo_description":14,"update_tm":29,"read_time":30},126580,687207017582,"Himbo","https://ap-avatar.wpscdn.com/davatar_994ba38a5ba835b3df7d355c54d3ed8d",8,"Research & Report","Tracking Machine Learning Bias Creep in Traditional and Online Lending Systems with Covariance Analysis","Machine Learning (ML) algorithms used in online banking guide consumer credit-card, auto-loan, and mortgage decisions, and may produce unfair outcomes toward certain groups. A common bias remedy removes sensitive attributes, yet they can remain encoded indirectly through correlated signals in the data. This paper proposes a covariance-analysis-based method to detect attributes that mimic sensitive information. Experiments on two credit datasets from traditional and online institutions show effective sensitive-attribute encapsulation and bias reduction while preserving overall model performance.","Open Research Online   \nThe Open University’s repository of research publications and other research outputs  \nTracking Machine Learning Bias Creep in Traditional and Online Lending Systems with Covariance Analysis  \nConference or Workshop Item  \nHow to cite:  \nPavón Pérez, Ángel; Fernandez, Miriam; Al-Madfai, Hasan; Burel, Grégoire and Alani, Harith (2023) . Tracking Machine Learning Bias Creep in Traditional and Online Lending Systems with Covariance Analysis. In: Proceedings of the 15th ACM Web Science Conference 2023, Association for Computing Machinery, New York, NY, United States pp. 184–195.  \nFor guidance on citations see FAQs.  \n􀀍c 2023 The Authors  \n[https://creativecommons.org/licenses/by-nc-nd/4.0/](https://creativecommons.org/licenses/by-nc-nd/4.0/)  \nVersion: Version of Record  \nLink(s) to article on publisher’s website:  \n[http://dx.doi.org/doi:10.1145/3578503.3583605](http://dx.doi.org/doi:10.1145/3578503.3583605)  \nCopyright and Moral Rights for the articles on this site are retained by the individual authors and/or other copyright owners. For more information on Open Research Online’s data policy on reuse of materials please consult the policies page.  \n[oro.open.ac.uk](oro.open.ac.uk)  \nTracking Machine Learning Bias Creep in Traditional and Online Lending Systems with Covariance Analysis  \nÁngel Pavón Pérez  \nThe Open University Milton Keynes, United Kingdom  \nMiriam Fernandez  \nThe Open University Milton Keynes, United Kingdom  \nHasan Al-Madfai Visa Europe  \nLondon, United Kingdom  \nGrégoire Burel  \nThe Open University Milton Keynes, United Kingdom  \nHarith Alani  \nThe Open University Milton Keynes, United Kingdom  \nABSTRACT  \nMachine Learning (ML) algorithms are embedded within online banking services, proposing decisions about consumers’ credit cards, car loans, and mortgages. These algorithms are sometimes biased, resulting in unfair decisions toward certain groups. One common approach for addressing such bias is simply dropping the sensitive attributes from the training data (e.g. gender) . However, sensitive attributes can indirectly be represented by other attributes in the data (e.g. maternity leave taken) . This paper addresses the problem of identifying attributes that can mimic sensitive attributes by proposing a new approach based on covariance analysis. Our evaluation conducted on two different credit datasets, extracted from a traditional and an online banking institution respectively, shows how our approach: (i) effectively identifies the attributes from the data that encapsulate sensitive information and,(ii) leads to the reduction of biases in ML models, while maintaining their overall performance.  \nCCS CONCEPTS  \n• Computing methodologies → Machine learning; Featureselection; • Information systems → World Wide Web.  \nKEYWORDS  \nMachine Learning, Financial Services, Bias in data, Bias identification, Bias mitigation  \nACM Reference Format:  \nÁngel Pavón Pérez, Miriam Fernandez, Hasan Al-Madfai, Grégoire Burel, and HarithAlani. 2023. Tracking Machine Learning Bias Creep in Traditional and Online Lending Systems with Covariance Analysis. In 15th ACM Web Science Conference 2023 (WebSci ’23), April 30–May 01, 2023, Austin, TX, USA. ACM, New York, NY, USA, 12 pages. [https://doi.org/10.1145/3578503.3583605](https://doi.org/10.1145/3578503.3583605)  \n1 INTRODUCTION  \nAutomatic decision-making based on large amounts of data ingested by Machine Learning (ML) models has become increasingly present  \nPermission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. Copyrights for components of this work owned by others than the author(s) must be honored. Abstracting with credit is permitted. To copy otherwise, or republish, to post on servers or to redistribute to lists, requires prior sp","cbCairBGoEUOVDaG","https://ap.wps.com/l/cbCairBGoEUOVDaG","pdf",1652805,2,1,13,"English","en",105,"# Introduction\n## Problem of algorithmic bias in lending decisions\n## Bias can persist without explicit sensitive attributes\n# Approach\n## Covariance analysis for bias-encapsulating attributes\n# Evaluation\n## Results on traditional and online credit datasets","[{\"question\":\"Why can ML bias persist even when sensitive attributes like gender are removed from training data?\",\"answer\":\"Sensitive traits may still be indirectly represented by other correlated attributes in the dataset. For example, parental or maternity leave can encode gender-related information.\"},{\"question\":\"What does the proposed covariance-analysis approach aim to identify?\",\"answer\":\"It identifies attributes that mimic sensitive attributes by detecting how information related to sensitive traits is captured in the data through covariance patterns.\"},{\"question\":\"How is the approach evaluated and what outcomes are reported?\",\"answer\":\"Evaluation uses two credit datasets from traditional and online lending institutions, showing that the method captures sensitive information encapsulated in data and reduces bias in ML models while maintaining overall performance.\"}]","Tracking Machine Learning Bias Creep in Traditional and Online Lending Systems with Covariance Analysis | PDF",1785933460,33,{"code":4,"msg":32,"data":33},"ok",{"site_id":25,"language":24,"slug":34,"title":13,"keywords":35,"description":14,"schema_data":36,"social_meta":87,"head_meta":89,"extra_data":91,"updated_unix":29},"tracking-machine-learning-bias-creep-in-traditional-and-online-lending-systems-with-covariance-analysis","",{"@graph":37,"@context":86},[38,54,69],{"@type":39,"itemListElement":40},"BreadcrumbList",[41,45,48,51],{"item":42,"name":43,"@type":44,"position":21},"https://docshare.wps.com","Home","ListItem",{"item":46,"name":47,"@type":44,"position":20},"https://docshare.wps.com/document/","Document",{"item":49,"name":12,"@type":44,"position":50},"https://docshare.wps.com/document/research-report/",3,{"item":52,"name":13,"@type":44,"position":53},"https://docshare.wps.com/document/tracking-machine-learning-bias-creep-in-traditional-and-online-lending-systems-with-covariance-analysis/126580/",4,{"url":52,"name":13,"@type":55,"author":56,"headline":13,"publisher":58,"fileFormat":61,"inLanguage":24,"description":14,"dateModified":62,"datePublished":63,"encodingFormat":61,"isAccessibleForFree":64,"interactionStatistic":65},"DigitalDocument",{"name":9,"@type":57},"Person",{"url":42,"name":59,"@type":60},"DocShare","Organization","application/pdf","2026-08-28","2026-08-05",true,{"@type":66,"interactionType":67,"userInteractionCount":20},"InteractionCounter",{"@type":68},"ViewAction",{"@type":70,"mainEntity":71},"FAQPage",[72,78,82],{"name":73,"@type":74,"acceptedAnswer":75},"Why can ML bias persist even when sensitive attributes like gender are removed from training data?","Question",{"text":76,"@type":77},"Sensitive traits may still be indirectly represented by other correlated attributes in the dataset. For example, parental or maternity leave can encode gender-related information.","Answer",{"name":79,"@type":74,"acceptedAnswer":80},"What does the proposed covariance-analysis approach aim to identify?",{"text":81,"@type":77},"It identifies attributes that mimic sensitive attributes by detecting how information related to sensitive traits is captured in the data through covariance patterns.",{"name":83,"@type":74,"acceptedAnswer":84},"How is the approach evaluated and what outcomes are reported?",{"text":85,"@type":77},"Evaluation uses two credit datasets from traditional and online lending institutions, showing that the method captures sensitive information encapsulated in data and reduces bias in ML models while maintaining overall performance.","https://schema.org",{"og:url":52,"og:type":88,"og:title":13,"og:site_name":59,"og:description":14},"article",{"robots":90,"canonical":52},"index,follow",{"doc_id":7,"site_id":25},{"code":4,"msg":5,"data":93},[94,98,102,106,111,116,121,124,129,132,136],{"id":21,"doc_module":4,"doc_module_name":47,"category_name":95,"show_sort_weight":96,"slug":97},"Story & Novel",90,"story-novel",{"id":20,"doc_module":4,"doc_module_name":47,"category_name":99,"show_sort_weight":100,"slug":101},"Literature",80,"literature",{"id":53,"doc_module":4,"doc_module_name":47,"category_name":103,"show_sort_weight":104,"slug":105},"Exam",70,"exam",{"id":107,"doc_module":4,"doc_module_name":47,"category_name":108,"show_sort_weight":109,"slug":110},5,"Comic",60,"comic",{"id":112,"doc_module":4,"doc_module_name":47,"category_name":113,"show_sort_weight":114,"slug":115},6,"Technology",50,"technology",{"id":117,"doc_module":4,"doc_module_name":47,"category_name":118,"show_sort_weight":119,"slug":120},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":47,"category_name":12,"show_sort_weight":122,"slug":123},30,"research-report",{"id":125,"doc_module":4,"doc_module_name":47,"category_name":126,"show_sort_weight":127,"slug":128},9,"Religion & Spirituality",20,"religion-spirituality",{"id":127,"doc_module":4,"doc_module_name":47,"category_name":130,"show_sort_weight":127,"slug":131},"World Cup","world-cup",{"id":133,"doc_module":4,"doc_module_name":47,"category_name":134,"show_sort_weight":133,"slug":135},10,"Lifestyle","lifestyle",{"id":137,"doc_module":4,"doc_module_name":47,"category_name":138,"show_sort_weight":107,"slug":139},19,"General","general"]