BAYESIAN PARAMETRIC PREDICTIVE MODELING OF GROUP CLAIMS INSURANCE
Bayesian methods combining two sources of information about the parameters of a statistical model. The combination of sample information (likelihood function) and prior information (prior distribution) will generate posterior information (posterior disribution). In this thesis, the posterior probabi...
Saved in:
| Main Author: | |
|---|---|
| Format: | Theses |
| Language: | Indonesia |
| Online Access: | https://digilib.itb.ac.id/gdl/view/22223 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Institut Teknologi Bandung |
| Language: | Indonesia |
| Summary: | Bayesian methods combining two sources of information about the parameters of a statistical model. The combination of sample information (likelihood function) and prior information (prior distribution) will generate posterior information (posterior disribution). In this thesis, the posterior probability function that has been generated, is used to compute predictive probability and expectation of severity for new group. The data are severity of group insurance where the zero claims probability is positive in each groups. Furthermore, there will be formed new groups that combine the characteristics of the exist groups to compare the probability of a risk in each new groups. Analytically, posterior distribution is difficult to determine. Therefore, we use computational program through Monte Carlo simulation, known as simulated Markov Chain Monte Carlo (MCMC) with Metropolis-Hasting algorithm. In this study, Metropolis-Hasting algorithm will be used to estimate the parameters of the new group insurance. The result is groups which consist of at least a group with high severity generate a new group with high severity expectation. |
|---|