Akademik Veri Yönetim Sistemi
IV. International Applied Statistics Conference (UYIK – 2023), Sarajevo, Bosna-Hersek, 26 - 29 Eylül 2023, ss.1
Traffic insurance is one of the most fundamental legal obligations for all vehicle owners. Insurance companies are obligated to provide compensation payments such as material, death and invalidity compensation, under the umbrella of traffic insurance following the occurrence of a claim. In this study, insurance compensation payments which are pivotal expenses for insurance companies are modeled using multivariate statistical methods. The assumption of gamma distribution for logarithmic compensation payments are used with both regression-based and copula-based multivariate methods. Regression-based models are considered as bivariate gamma regression which is established using the bivariate gamma distribution, and the mixture of bivariate gamma regressions with the mixture of experts, one of the machine learning techniques. Copula-based multivariate models are discussed as bivariate copula regression and finite mixture of copula regression models which are designed using Gumbel and Frank copula functions. A comparative analysis is conducted using insurance compensation payments for Turkish motor vehicles compulsory third party liability insurance between 2018 and 2022 by mvClaim package in R. It is noticed that, the mixture of copula-based models is more suitable for the multivariate modeling of insurance compensation payments.