STRIKE: Additive Feature-Group-Aware Stacking Framework for Credit Default Prediction

2026-04-19

AI Summary

This study addresses the challenges posed by high-dimensional, heterogeneous, and noisy credit data, which often lead traditional monolithic models to overfit and exhibit poor robustness under distributional shifts. To mitigate these issues, the authors propose a feature-group-aware stacking framework that partitions the feature space into semantically coherent groups, trains dedicated learners on each group, and aggregates their predictions via a meta-learner to construct a modular and interpretable additive risk model. Rather than increasing model complexity alone, this approach enhances performance through meaningful feature decomposition. Empirical evaluation on three real-world credit risk datasets demonstrates that the proposed method achieves significantly higher AUC-ROC scores than strong baseline tree-based models and conventional stacking approaches, confirming its superior generalization ability and stability.

Technology Category

• Machine Learning: Ensemble Methods
• Reasoning under Uncertainty: Relational Probabilistic Models
• Natural Language Processing: Interpretability, Analysis, and Evaluation of NLP Models

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalization

Abstract

Credit risk default prediction remains a cornerstone of risk management in the financial industry. The task involves estimating the likelihood that a borrower will fail to meet debt obligations, an objective critical for lending decisions, portfolio optimization, and regulatory compliance. Traditional machine learning models such as logistic regression and tree-based ensembles are widely adopted for their interpretability and strong empirical performance. However, modern credit datasets are high-dimensional, heterogeneous, and noisy, increasing overfitting risk in monolithic models and reducing robustness under distributional shift. We introduce STRIKE (Stacking via Targeted Representations of Isolated Knowledge Extractors), a feature-group-aware stacking framework for structured tabular credit risk data. Rather than training a single monolithic model on the complete dataset, STRIKE partitions the feature space into semantically coherent groups and trains independent learners within each group. This decomposition is motivated by an additive perspective on risk modeling, where distinct feature sources contribute complementary evidence that can be combined through a structured aggregation. The resulting group-specific predictions are integrated through a meta-learner that aggregates signals while maintaining robustness and modularity. We evaluate STRIKE on three real-world datasets spanning corporate bankruptcy and consumer lending scenarios. Across all settings, STRIKE consistently outperforms strong tree-based baselines and conventional stacking approaches in terms of AUC-ROC. Ablation studies confirm that performance gains stem from meaningful feature decomposition rather than increased model complexity. Our findings demonstrate that STRIKE is a stable, scalable, and interpretable framework for credit risk default prediction tasks.

Problem

• credit default prediction
• high-dimensional data
• distributional shift
• overfitting
• feature heterogeneity

Innovation

• feature-group-aware
• stacking framework
• credit default prediction
• additive risk modeling
• modular ensemble learning