How Data Quality Affects Machine Learning Models for Credit Risk Assessment

Overview

Andrea Maurino's research investigates the impact of data quality on machine learning models for credit risk assessment. The study uses controlled data corruption to evaluate model robustness.

Key Insights

• Data Quality Impact: Data issues like missing values and noisy attributes significantly affect model accuracy.
• Model Robustness: Models such as Random Forest and SVM show varying robustness depending on data degradation severity.
• Practical Framework: The study provides a framework for enhancing data pipeline robustness.

Why It Matters

Data quality is crucial for accurate credit risk assessment, impacting financial decision-making and risk management.

Actionable Implications

• Evaluate and improve data quality in credit risk models.
• Use the provided framework to test model robustness against data issues.
• Incorporate data quality checks in data pipelines.