rating-factors

GLM tooling for insurance pricing — nested GLM embeddings, R2VF factor level clustering, territory banding, SKATER

R2VF factor level clustering for insurance GLMs — automated categorical banding with Poisson/Gamma/Tweedie, exposure offsets, monotonicity, BIC lambda selection (156 tests)

Automated GLM interaction detection for insurance pricing — CANN, NID scoring, SHAP interaction values

SHAP-based rating relativities from GBM models for insurance pricing — extract GLM-style factors from CatBoost, Polars

Extract multiplicative rating relativities from GBMs using SHAP values. Built for insurance pricing.