Gregorius Reynaldi Pratama

Accurate Ensembles, Fragile Narratives: Multi-Scale Stacking and a Fidelity Audit of LLM-Generated Explanations for Credit Risk

This work builds a multi-scale stacking ensemble fusing four differently regularised gradient-boosting learners with a residual network through a neural meta-learner trained on out-of-fold predictions that fails in a way prompt engineering alone does not fix credit scoring.