THE ROLE OF BOOTSTRAP RESAMPLING METHOD ON FIRE INSURANCE DATA
The use of big data has become trendy in the business world to collect customer information from every transaction that occurs. In the insurance field, big data can be used to estimate premiums. In this paper, the writer wants to apply the bootstrap resampling method to fire insurance that is quite...
Saved in:
| Main Author: | |
|---|---|
| Format: | Final Project |
| Language: | Indonesia |
| Online Access: | https://digilib.itb.ac.id/gdl/view/47714 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Institut Teknologi Bandung |
| Language: | Indonesia |
| Summary: | The use of big data has become trendy in the business world to collect customer information from every transaction that occurs. In the insurance field, big data can be used to estimate premiums. In this paper, the writer wants to apply the bootstrap resampling method to fire insurance that is quite large in size. This data processing is carried out based on samples taken at random systematically to get the sample size to be used is 1%, 5%, and 10%. Furthermore, these samples will be processed to calculate insurance premiums using the principle of expected value. In this principle, parameter such as expectation and variance of the total claims are required. These two parameters will be estimated through bootstrap samples. Bootstrap sample is a sample derived from the results of resampling with the return of the observed sample as many times as replication B. After the estimate is obtained, it will be compared between the three samples which produce the best bootstrap estimate. The criteria to be considered in determining the best bootstrap estimates are based on the smallest standard error, the smallest percentile confidence interval, and the closest to normal bootstrap histogram. The amount of premium generated from the bootstrap sample is expected to be close to the amount of premium generated from population data. |
|---|