kockázatkalibrálás
kockázatkalibrálás refers to the systematic adjustment of risk estimates to better reflect observed outcomes. In banking and insurance, it is used to align probability‑of‑default or loss‑given‑default models with historical experience. The process typically involves statistical tests such as the Hosmer‑Lemeshow test, which compares predicted event frequencies with actual frequencies across deciles of risk, and calibration plots that display predicted versus observed default rates. When a model shows poor calibration, adjustments are made—often by reweighting or recalibrating parameters—so that risk scores become more accurate.
The concept emerged alongside the development of credit scoring and rating systems in the 1980s and has
Practical implementation varies: financial institutions may build separate calibration models or apply margin adjustments to base