DETERMINING AUTO INSURANCE PREMIUM USING GENERALIZED GEO-ADDITIVE MODEL
Motor vehicle insurance is an insurance policy that provides coverage for losses that occur to motor vehicles. In order to obtain financial protection against losses that occur to their motor vehicles, vehicle owners need to pay a premium to the insurance company. The amount of premium paid cannot b...
Saved in:
| Main Author: | |
|---|---|
| Format: | Final Project |
| Language: | Indonesia |
| Online Access: | https://digilib.itb.ac.id/gdl/view/73001 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Institut Teknologi Bandung |
| Language: | Indonesia |
| Summary: | Motor vehicle insurance is an insurance policy that provides coverage for losses that occur to motor vehicles. In order to obtain financial protection against losses that occur to their motor vehicles, vehicle owners need to pay a premium to the insurance company. The amount of premium paid cannot be the same for all policyholders because each policy has different characteristics of the vehicle. The determination of the premium needs to consider the risk factors of the insured and the policyholders. In addition, geographical aspects also have a significant influence on the frequency of motor accidents or cases of motor vehicle loss. By using motor vehicle insurance claim data from a general insurance company in Indonesia, the influence of risk factors, namely vehicle category, usage type, and region, on the frequency and severity of claims will be examined. The multiplication of the claim frequency expectation and claim severity will result in a pure premium for each policy. To model the relationship between the predictor variables and the response variable while also considering the geographical aspect or region, a Generalized Geo-additive Model (GGAM) will be used. GGAM is an extension of the Generalized Additive Model (GAM) with additional structured and unstructured spatial random effects in the model. In this thesis, a model will be developed using data from policyholders in the year 2016, with a total of 737,376 policies and 5,382 claims. The results show that claim severity is influenced by vehicle category, usage type, and region, while claim frequency is influenced by vehicle category. Additionally, cluster analysis using the k-means method will be used to determine regional clusters consisting of provinces in Indonesia. Provinces within the same cluster will have the same premium amount. |
|---|