Advanced financial risk forecasting using enhanced kernel-based extreme learning machines: Tackling challenges in bankruptcy problem

Accurately forecasting financial risk, particularly in bankruptcy forecasting, remains a significant challenge due to the non-linear and complex nature of economic indicators. Conventional statistical techniques often fail to capture the intricate relationships within financial datasets. This research addresses these shortcomings using a Kernel-based Extreme Learning Machine (KELM). KELM stands out for its capability to discern complex patterns, yet the model’s full potential is often underutilized due to the intricate hyperparameter tuning process. We identify a critical research gap: the need for sophisticated algorithms that can effectively fine-tune KELM tailored for financial risk forecasting. To bridge this gap, our study develops the Objective-based Survival Individual Enhancement Chimp Optimization Algorithm (OBSIECOA). This innovative algorithm elevates hyperparameter optimization by bolstering weaker predictions and preserving diversity among the strongest. It strategically reallocates the elite candidates during iterative cycles, enabling an extensive and dynamic exploration of the hyperparameter space. Our methodology emphasizes the importance of precision in financial forecasting models, and the OBSIECOA’s contribution is quantitatively assessed through rigorous evaluation against several KELM variants: standard KELM, KELM-DBOA, KELM-MOA, KELM-NFVOA, and KELM-WOBCOA. The assessment uses two real-world financial datasets, the Japanese Dataset (JPNBDS) and the Wieslaw dataset, with performance metrics including RMSE, NSEF, and bias. The KELM-OBSIECOA framework significantly improves forecasting accuracy over traditional models, confirming its efficacy as a superior early warning system for financial distress. Our findings underscore the necessity for and effectiveness of KELM approaches in overcoming the inherent challenges of financial risk prediction. The research advocates for integrating such advanced models in financial analysis, highlighting their strategic importance in anticipating and managing economic uncertainties.