Generalized Linear Model for Compound Model with Dependency between The Primary and The Secondary Random Variables
Generalized linear models may be used to determine an insurance pure premium, especially in a non-life insurance business. If a compound model is used to model an aggregate loss, then usually it is assumed that the number of claims (claims frequency) and the amount of claims (claims severity) are...
Saved in:
| Main Author: | |
|---|---|
| Format: | Theses |
| Language: | Indonesia |
| Online Access: | https://digilib.itb.ac.id/gdl/view/36139 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Institut Teknologi Bandung |
| Language: | Indonesia |
| Summary: | Generalized linear models may be used to determine an insurance pure premium,
especially in a non-life insurance business. If a compound model is used to model
an aggregate loss, then usually it is assumed that the number of claims (claims
frequency) and the amount of claims (claims severity) are independent. Hence,
the pure premium is the product of the marginal mean frequency and the marginal
mean severity. However, ther are cases where the claims frequency and the claims
severity are not mutually independent. This Thesis discusses modelling an aggregate
loss random variable using a compound model with a generalized linear
model approach when the claims frequency and the claims severity are not mutually
independent. In this Thesis, the claims frequency is assumed to follow a
poisson distribution; and the claims severity is assumed to follow a gamma distribution.
Each of the claims frequency and severity is separately modelled using
a generalized linear model with the log link function. For a case study, a car
insurnace data is analyzed. |
|---|