How Data Quality Affects Machine Learning Models for Credit Risk Assessment
Andrea Maurino
TL;DR
The paper investigates how data quality affects credit risk ML predictions by applying controlled data degradations with the PuckTrick library to an open-source credit risk dataset across 10 classifiers, evaluated via $F1$ on clean test data. It finds that certain errors can enhance predictive performance (up to around 17% in $F1$) while others degrade results, with effects depending on the corrupted feature and the classifier. The study introduces a modular, reproducible methodology for injecting data imperfections and assessing robustness, offering practical guidelines for strengthening data pipelines in financial settings. Overall, the work highlights the importance of data quality management and labeling accuracy in real-world credit risk applications and provides a framework for broader data-centric AI investigations.
Abstract
Machine Learning (ML) models are being increasingly employed for credit risk evaluation, with their effectiveness largely hinging on the quality of the input data. In this paper we investigate the impact of several data quality issues, including missing values, noisy attributes, outliers, and label errors, on the predictive accuracy of the machine learning model used in credit risk assessment. Utilizing an open-source dataset, we introduce controlled data corruption using the Pucktrick library to assess the robustness of 10 frequently used models like Random Forest, SVM, and Logistic Regression and so on. Our experiments show significant differences in model robustness based on the nature and severity of the data degradation. Moreover, the proposed methodology and accompanying tools offer practical support for practitioners seeking to enhance data pipeline robustness, and provide researchers with a flexible framework for further experimentation in data-centric AI contexts.