Evaluating Predictive Models for Precision in Insurance Pricing

In the insurance and healthcare sectors, accurately forecasting medical insurance premiums is essential. This study compares a number of regression models, to estimate insurance costs based on demographic and lifestyle data. We applied feature scaling to improve model consistency. Grid search was used to adjust the Light GBM hyperparameters, and Light GBM was used as the meta-learner in the stacking ensemble. Each model was evaluated using the coefficient of determination The Stacking Regressor demonstrated the efficiency of model ensembling by outperforming all other models with the highest coefficient of determination score. This Comparative approach highlights ensemble and hybrid approaches for accurate cost forecasting in insurance applications and clarifies the predictive capabilities of different algorithms.