IMPLEMENTATION OF FUZZY C-MEANS ALGORITHM AND FUZZY INFERENCE SYSTEM ON INSURANCE DATA
The cash inflow of an insurance company is mainly coming from the premiums paid. In general, the premiums set by the company are higher than the premiums calculated by the equivalence principle. As the main source of cash, premiums need to be well-managed to increase the company’s income from in...
Saved in:
| Main Author: | |
|---|---|
| Format: | Final Project |
| Language: | Indonesia |
| Online Access: | https://digilib.itb.ac.id/gdl/view/65109 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Institut Teknologi Bandung |
| Language: | Indonesia |
| Summary: | The cash inflow of an insurance company is mainly coming from the
premiums paid. In general, the premiums set by the company are higher than the
premiums calculated by the equivalence principle. As the main source of cash,
premiums need to be well-managed to increase the company’s income from
investments while still be able to cover future risks. However, risk related
information from secondary data are commonly incomplete. Therefore, the fuzzy
concept is considered to analyze the incomplete information. The fuzzy concept is
a logical concept with a vague (fuzzy) truth value, hence able to describe
incomplete information. The fuzzy concepts used in this study consists of the fuzzy
c-means algorithm as a data grouping method and the fuzzy inference system as a
logic gate in taking decisions. The policyholders’ data are grouped by age into four
categories, which are young, young adult, adult, and old. While based on premium
price level (from the policy’s description), data is grouped into four groups, which
are cheap, fair, expensive, and extremely expensive. Both variables are used as an
input to the fuzzy inference system to create an output of percentage premium to
save. Percentage premium to save is split into six categories, which are minimum,
low, moderate, moderate high, high, and maximum.
The fraction of the received premium can then be calculated by using the created
fuzzy inference system. The reserved premium is then accumulated, creating the
company’s reserves, while the rest of the fund is invested accordingly. Reserves are
then compared to the sum of valid policy values, that is the difference of expected
present value of benefits and premiums. The accumulated reserve only covered a
fraction of the valid policy value. This means the company will not be able to cover
future risks of the valid policies. Therefore, the company needs to allocate a part of
its investment funds to the reserves periodically so that the reserve can cover the
valid policy value. |
|---|