Analiza porównawcza modeli mieszanych szacowania stóp taryf w ubezpieczeniach majątkowych z wykorzystaniem kroswalidacji
Comparative analysis of mixed models for ratemaking in non-life insurance with k-fold cross-validation

Summary/Abstract: The paper presents the application of hierarchical generalized linear models (HGLM) in the process of a’priori ratemaking. As explained variable assumed the average value of claims paid (the continuous variable) and as explanatory variables - variables characterizing the subject and the person insured (nominal variables). The important computational problem in HGLM models is to determine the deviance which can be used as a criterion for the selection of the final form of the model. Thus, in order to simplify the first part of computing a’priori ratemaking (omitting the calculation of the deviance), k-fold cross-validation procedure was used and the estimated k-fold cross-validation prediction error was calculated. The case study was presented for a portfolio of motor insurance. The procedure was implemented in a computer program R.

• Details
• Contents