Summary of - Efficacies of Artificial Neural Networks Ushering Improvement in the Prediction of Extant Credit Risk Models
Document Type
Article
Abstract
Study Background: The study evaluates whether Machine Learning Techniques (ML) surpass Logistic Regression (LR) in predicting firm bankruptcy and whether including a proxy for uncertainty improves the model’s predictive power.
Research Goals and Hypotheses: The goal is to enhance credit risk prediction models using ML techniques and to assess the improvement when incorporating uncertainty proxies. The hypothesis is that ML techniques will outperform LR and that the predictive accuracy will increase with uncertainty proxies.
Methodological Approach: The research involves comparing the predictive power of ML techniques and LR for bankruptcy prediction, using a change in operating expenditure as a proxy for uncertainty.
Results and Discoveries: ML techniques showed superior predictive power over LR. Incorporating the proxy for uncertainty significantly improved the accuracy of bankruptcy predictions.
Publication Date: 2023
Recommended Citation: Aranha, M., & Bolar, K. (2023). Efficacies of artificial neural networks ushering improvement in the prediction of extant credit risk models. Cogent Economics and Finance, 11(1)
Publication Date
2023
Recommended Citation
Aranha, M and Bolar, K, "Summary of - Efficacies of Artificial Neural Networks Ushering Improvement in the Prediction of Extant Credit Risk Models" (2023). Open Access archive. 9171.
https://impressions.manipal.edu/open-access-archive/9171
