Customer Credit Worthiness in the Digital Age - A Management Approach to Machine Learning Application in Banking

Aug 5, 2026

Upload by Violet

This study investigates the shift in private banking from conventional creditworthiness assessment to advanced machine learning (ML) models. By synthesizing prior literature and case studies, it shows how ML models can combine alternative big data with advanced algorithms to improve default forecasting accuracy and support financial inclusion. The research emphasizes legal and ethical risks from alternative data processing, requiring compliance checks by banks and regulators, and calls for technical capacity to keep models transparent and interpretable, avoiding “black boxes.”

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages8

Reading Time20 min

AccessFree to download

Introduction

Digital transformation and first-principles change

Banking credit risk and creditworthiness scoring

Credit scoring with ML and big data in banking

Challenges: ethical/legal issues and discrimination risks

Vol. 06, No. 2 (2024) 665-672, doi: 10.24874/PES06.02.023 Proceedings on Engineering Sciences [www.pesjournal.net](www.pesjournal.net) CUSTOMER CREDIT WORTHINESS IN THE DIGITAL AGE: A MANAGEMENT APPROACH TO MACHINE LEARNING APPLICATION IN BANKING Henrik Manukyan 1 Suren Parsyan Received 14.06.2023. Received in revised form 07.10.2023. Accepted 19.11.2023. UDC – 004.85 Keywords: Risk Management, Machine Learning, Credit Scoring Models, Alternative Data, Ethical Considerations, Regulatory Oversight A B S T R A C T This study investigates the shift in private banking from conventional creditworthiness assessment to advanced machine learning (ML) models. Employing a synthesis technique, this study conducts a review of literature and case studies and highlights how ML models, through the integration of alternative big data and advanced algorithms, can enhance accuracy in forecasting customer defaults and contribute to financial inclusion. The research underscores legal and ethical concerns regarding alternative data processing, necessitating thorough compliance checks by banks and regulatory authorities. Furthermore, it underlines the necessity for banks and regulators to develop technical skills to ensure ML models remain transparent and understandable, avoiding the pitfalls of becoming “black boxes”. Future research is suggested to explore risk mitigation strategies based on its ML deployment approach, technical aspects of ML algorithms, and the impact of ML-based credit scoring on broader macro-financial linkages. © 2024 Published by Faculty of Engineering 1. INTRODUCTION In the rapidly evolving landscape of digital technologies companies are undergoing digital transformation. Digital transformation encompasses more than digitalization or digitization of documents (Schallmo & Williams, 2018) . In the academic literature, digital transformation transcends mere digitization, which typically involves converting objects into digital formats, often seen with paper documents. Digitalization, by contrast, refers to integrating business processes into the digital domain, while digital transformation implies a broader range of changes within the organization. It encompasses not only the adoption of new technological tools but also significant alterations in work practices and organizational culture. In other words, it involves changes not only in the new technological tools used for performing the job but also in the logic and problems of the work process itself. This paradigm shift is rooted in the ―first principles‖(Brett, 2019) approach of digital transformation, which necessitates revisiting and reviewing existing operations and, where necessary, undertaking revolutionary changes. Digital transformation aims to maximize companies‘ efficiency and strives for data-driven decision-making. It is essential to distinguish between the changes that have in their core the ―first principles‖ approach from those following a ―design by analogy‖ approach, where business processes are modified, albeit non-disruptively, to improve existing procedures. In the digital era substantive, long-term value is anticipated from changes driven by the ―first principles‖ approach, potentially leading to a profound transformation in the very logic of the business processes. The banking sector, which occupies a prominent place in the financial system, comprises the central bank (sometimes referred to as the national bank), along with the public and private banks. The banking sector is also embracing digital technologies, and it often collaborates with technological companies to ensure a smooth transition into digital transformation. The primary goal of private banks (hereinafter banks) is the maximization of profit. Banks primarily engage in deposit intake and loan issuance to customers, with the latter encompassing credit risk – the risk associated with borrowers‘potential default on loans, when the borrowers fail torepay thei

FAQs

How does the study describe the transition in private banking toward machine learning models?

It describes a move from conventional creditworthiness assessment to advanced ML models that leverage alternative big data and advanced algorithms for better forecasting.

What is the expected benefit of using ML in credit scoring systems?

ML is expected to enhance predictive power and improve the accuracy of credit risk assessments, including default forecasting.

What risks and compliance concerns does the research highlight for ML-based creditworthiness models?

The study highlights legal and ethical concerns around alternative data processing, potential discrimination, and related reputation and regulatory compliance risks.

Aug 5, 2026

Upload by Violet