DETERMINATION OF PURE PREMIUM OF VEHICLE INSURANCE WITH GENERALIZED LINEAR MODEL

• Main Author: Nabila Shaliha, Anisa
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/74347
• Institution: Institut Teknologi Bandung

Currently, general insurance and life insurance have been very developed. One of the general insurances is vehicle insurance which is used by people who have private vehicles such as cars or motorbikes so that if one day a damage occurs, financial losses can be transferred to the insurance company. The premiums set by insurance companies for policyholders certainly vary depending on the risk factors of each policyholder. This research aims to determine a model for calculating vehicle premiums by considering the variables response of claim frequency and severity of claims which are influenced by predictor variables. The data used in this study is vehicle insurance data in the United States where data can be downloaded openly. Modeling is done by utilizing the Generalized Linear Model (GLM) and the models to be obtained are compared using the Likelihood ratio test and the Wald test in order to obtain the best model. The response variable from the claim frequency model follows the Poisson distribution, and the response variable from the claim severity model follows the Inverse Gaussian distribution. A pure premium are obtained by the multiplication of the expected frequency of claims with the expected severity of claims which are influenced by the variables of Coverage, Gender, Income, Location Code, Marital Status, Months Since Policy Inception, Number of Open Complaints, Policy, Claim Reason, Total Claim Amount, dan Vehicle Class.