Modeling Credit Risk: A Category Theory Perspective

Abstract

This paper proposes a conceptual modeling framework based on category theory that serves as a tool to study common structures underlying diverse approaches to modeling credit default that at first sight may appear to have nothing in common. The framework forms the basis for an entropy-based stacking model to address issues of inconsistency and bias in classification performance. Based on the Lending Club’s peer-to-peer loans dataset and Taiwanese credit card clients dataset, relative to individual base models, the proposed entropy-based stacking model provides more consistent performance across multiple data environments and less biased performance in terms of default classification. The process itself is agnostic to the base models selected and its performance superior, regardless of the models selected.

Kelly Criterion for Optimal Credit Allocation

“…In particular, the findings point to two key considerations for the design of an effective allocation system in practice. First, the equivalence concept proposed by category theory and detailed in an earlier paper (Tran et al 2021), provides a concrete basis for reasoning that combining several models would be a better way to construct optimal Kelly fractions. Confirmation of this proposition would require a level of analysis beyond the scope of the current paper.…”

“…Nine classifiers were selected as the base models for default and payment prediction: logistic regression, nearest neighbors, random forest, gradient boosted trees, decision tree, support vector machine, Markov, naive Bayes and neural network. Details of these base models are provided in Tran et al (2021).…”

“…See Tran et al (2021) for a discussion of the application of category theory to credit risk modeling.…”

An Ensemble Learning Based Strategy for Customer subdivision and Credit Risk Characterization

“…Stacking usually uses meta-learner as a combination method to re-learn the output of primary learners, improving the model's generalization ability by increasing the individual classifier's diversity and complementarity. Tran Cao Son et al [27] address inconsistency and bias in classification performance through the foundation of entropy-based stacking models. Xia Yufei et al [28] developed a heterogeneous stacking ensemble (HSE) approach and improved the loss-given default (LGD) forecasting in the P2P lending domain.…”