[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-126790-en":3,"doc-seo-126790-105":29,"detail-sidebar-cat-0-en-105":90},{"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":4,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":11,"language":21,"language_code":22,"site_id":23,"html_lang":22,"table_of_contents":24,"faqs":25,"seo_title":26,"seo_description":14,"update_tm":27,"read_time":28},126790,1099523882182,"Eliana","https://ap-avatar.wpscdn.com/davatar_6f874abed73319feea01a86fa6f0fab8",8,"Research & Report","How to Make Machine Learning Financial Recommendations More Fair: Theoretical Explanation","Machine learning is widely used to support financial decisions such as loan approvals and funding allocation. Although these systems often perform well overall, they can exhibit unexpected unfairness toward minority groups, sometimes exceeding the bias present in the training data. Prior empirical work showed that hyperparameter choices can influence fairness, with only a limited subset consistently improving it, notably feature count and training iterations. This report develops a theoretical explanation for those observed results.","University of Texas at El Paso  \nScholarWorks@UTEP  \n\n| Departmental Technical Reports (CS) | Computer Science |\n| --- | --- |\n| 10-1-2023\u003Cbr>How to Make Machine Learning Financial Recommendations More Fair: Theoretical Explanation\u003Cbr>Tho M. Nguyen\u003Cbr>Ho-Chi-Minh City Open University, [tho.nm@ou.edu.vn](tho.nm@ou.edu.vn)\u003Cbr>Saeid Tizpaz-Niari\u003Cbr>The University of Texas at El Paso, [saeid@utep.edu](saeid@utep.edu)\u003Cbr>Vladik Kreinovich\u003Cbr>The University of Texas at El Paso, [vladik@utep.edu](vladik@utep.edu)\u003Cbr>Follow this and additional works at: [https://scholarworks.utep.edu/cs_techrep](https://scholarworks.utep.edu/cs_techrep)\u003Cbr> Part of the Computer Sciences Commons, and the Mathematics Commons\u003Cbr>Comments:\u003Cbr>Technical Report: UTEP-CS-23-57 |  |\n\nRecommended Citation  \nNguyen, Tho M.; Tizpaz-Niari, Saeid; and Kreinovich, Vladik, \"How to Make Machine Learning Financial Recommendations More Fair: Theoretical Explanation\" (2023) . Departmental Technical Reports (CS). 1842.  \n[https://scholarworks.utep.edu/cs_techrep/1842](https://scholarworks.utep.edu/cs_techrep/1842)  \nThis Article is brought to you for free and open access by the Computer Science at ScholarWorks@UTEP. It has been accepted for inclusion in Departmental Technical Reports (CS) by an authorized administrator of ScholarWorks@UTEP. For more information, please contact [lweber@utep.edu](lweber@utep.edu).  \nHow to Make Machine Learning Financial Recommendations More Fair: Theoretical Explanation of Empirical Results  \nTho M. Nguyen, Saeid Tizpaz-Niari, and Vladik Kreinovich  \nAbstract Machine learning has been actively and successfully used to make financial decisions. In general, these systems work reasonably well. However, in some cases, these systems show unexpected bias towards minority groups – the bias that is sometime much larger than the bias in the data on which they were trained. A recent paper analyzed whether a proper selection of hyperparameters can decrease this bias. It turned out that while the selection of hyperparameters indeed affect the system’s fairness, only a few of the hyperparameters lead to consistent improvement of fairness: the number of features used for training and the number of training iterations. In this paper, we provide a theoretical explanation for these empirical results.  \n1 Formulation of the General Problem  \nMachine learning have become ubiquitous in decision making. In many practical situations, we need to make a recommendation. For example, if a person applies fora loan, the bank needs to decide whether providing this loan is worth a risk, and if yes, shall the bank give the full requested amount or a smaller amount. Similar decisions need to be made in other financial situations: e.g.  \n• when a start-up company applies for funding, or  \n• when a well-established company wants the bank to finance a project.  \nTho M. Nguyen  \nFaculty of Banking and Finance, Ho Chi Minh City Open University, Ho Chi Minh City, Vietnam, e-mail: [tho.nm@ou.edu.vn](tho.nm@ou.edu.vn)  \nSaeid Tizpaz-Niari and Vladik Kreinovich  \nDepartment of Computer Science,University of Texas at El Paso  \n500 W. University, El Paso, Texas 79968, USA, e-mail: [saeid@utep.edu](saeid@utep.edu), [vladik@utep.edu](vladik@utep.edu)  \n2 Tho M. Nguyen, Saeid Tizpaz-Niari, and Vladik Kreinovich  \nIn all such situations, there is no ready formulas for making a decision. So, we have to rely on specialists. Some specialists are more skilled, some are less skilled. It would therefore be nice to incorporate the expertise of skilled experts into a computer system – this would help less experienced decision makers make good decisions.  \nIn some cases, experienced experts can formulate their decision making process in terms of precise rules, but in financial domain, such situations are rare. In many cases, the decisions of experienced experts are motivated largely by their intuition, intuition that they cannot describe in precise terms.  \nSince we do not have the rules, all ","cbCailFtzxjppFiq","https://ap.wps.com/l/cbCailFtzxjppFiq","pdf",254211,1,"English","en",105,"# Abstract\n## Formulation of the General Problem\n## Problem Background: Decision Support and Expertise\n## Machine Learning and Regression Framing\n## Main Problem: Insufficient Fairness","[{\"question\":\"Why can machine learning financial recommendation systems be unfair even when they work well on average?\",\"answer\":\"They may produce acceptable average performance but still fail to be fair for applicants from certain minority groups, where results can become substantially unjust.\"},{\"question\":\"Which hyperparameters were found to consistently improve fairness in related empirical studies?\",\"answer\":\"The report notes that fairness improvements were consistent only for a few hyperparameters, particularly the number of features used for training and the number of training iterations.\"},{\"question\":\"What is the purpose of this paper in relation to prior empirical findings?\",\"answer\":\"It provides a theoretical explanation for why those specific hyperparameter choices lead to the observed fairness changes in machine learning financial recommendations.\"}]","How to Make Machine Learning Financial Recommendations More Fair: Theoretical Explanation | PDF",1785934800,20,{"code":4,"msg":30,"data":31},"ok",{"site_id":23,"language":22,"slug":32,"title":13,"keywords":33,"description":14,"schema_data":34,"social_meta":85,"head_meta":87,"extra_data":89,"updated_unix":27},"how-to-make-machine-learning-financial-recommendations-more-fair-theoretical-explanation","",{"@graph":35,"@context":84},[36,53,67],{"@type":37,"itemListElement":38},"BreadcrumbList",[39,43,47,50],{"item":40,"name":41,"@type":42,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":44,"name":45,"@type":42,"position":46},"https://docshare.wps.com/document/","Document",2,{"item":48,"name":12,"@type":42,"position":49},"https://docshare.wps.com/document/research-report/",3,{"item":51,"name":13,"@type":42,"position":52},"https://docshare.wps.com/document/how-to-make-machine-learning-financial-recommendations-more-fair-theoretical-explanation/126790/",4,{"url":51,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":22,"description":14,"dateModified":61,"datePublished":61,"encodingFormat":60,"isAccessibleForFree":62,"interactionStatistic":63},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":40,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-08-05",true,{"@type":64,"interactionType":65,"userInteractionCount":4},"InteractionCounter",{"@type":66},"ViewAction",{"@type":68,"mainEntity":69},"FAQPage",[70,76,80],{"name":71,"@type":72,"acceptedAnswer":73},"Why can machine learning financial recommendation systems be unfair even when they work well on average?","Question",{"text":74,"@type":75},"They may produce acceptable average performance but still fail to be fair for applicants from certain minority groups, where results can become substantially unjust.","Answer",{"name":77,"@type":72,"acceptedAnswer":78},"Which hyperparameters were found to consistently improve fairness in related empirical studies?",{"text":79,"@type":75},"The report notes that fairness improvements were consistent only for a few hyperparameters, particularly the number of features used for training and the number of training iterations.",{"name":81,"@type":72,"acceptedAnswer":82},"What is the purpose of this paper in relation to prior empirical findings?",{"text":83,"@type":75},"It provides a theoretical explanation for why those specific hyperparameter choices lead to the observed fairness changes in machine learning financial recommendations.","https://schema.org",{"og:url":51,"og:type":86,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":88,"canonical":51},"index,follow",{"doc_id":7,"site_id":23},{"code":4,"msg":5,"data":91},[92,96,100,104,109,114,119,122,126,129,133],{"id":20,"doc_module":4,"doc_module_name":45,"category_name":93,"show_sort_weight":94,"slug":95},"Story & Novel",90,"story-novel",{"id":46,"doc_module":4,"doc_module_name":45,"category_name":97,"show_sort_weight":98,"slug":99},"Literature",80,"literature",{"id":52,"doc_module":4,"doc_module_name":45,"category_name":101,"show_sort_weight":102,"slug":103},"Exam",70,"exam",{"id":105,"doc_module":4,"doc_module_name":45,"category_name":106,"show_sort_weight":107,"slug":108},5,"Comic",60,"comic",{"id":110,"doc_module":4,"doc_module_name":45,"category_name":111,"show_sort_weight":112,"slug":113},6,"Technology",50,"technology",{"id":115,"doc_module":4,"doc_module_name":45,"category_name":116,"show_sort_weight":117,"slug":118},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":45,"category_name":12,"show_sort_weight":120,"slug":121},30,"research-report",{"id":123,"doc_module":4,"doc_module_name":45,"category_name":124,"show_sort_weight":28,"slug":125},9,"Religion & Spirituality","religion-spirituality",{"id":28,"doc_module":4,"doc_module_name":45,"category_name":127,"show_sort_weight":28,"slug":128},"World Cup","world-cup",{"id":130,"doc_module":4,"doc_module_name":45,"category_name":131,"show_sort_weight":130,"slug":132},10,"Lifestyle","lifestyle",{"id":134,"doc_module":4,"doc_module_name":45,"category_name":135,"show_sort_weight":105,"slug":136},19,"General","general"]