Kosasih, Michelle (2025) Perhitungan premi asuransi kendaraan bermotor menggunakan generalized linear model dengan distribusi tweedie dan gradient tree-boosted tweedie model = Modeling motor vehicle insurance premiums using generalized linear model with tweedie distribution and gradient tree-boosted tweedie model. Bachelor thesis, Universitas Pelita Harapan.
![Title [thumbnail of Title]](http://repository.uph.edu/style/images/fileicons/text.png)
Title.pdf
Restricted to Registered users only
Available under License Creative Commons Attribution Non-commercial Share Alike.
Download (184kB)
![Abstract [thumbnail of Abstract]](http://repository.uph.edu/style/images/fileicons/text.png)
Abstract.pdf
Restricted to Registered users only
Available under License Creative Commons Attribution Non-commercial Share Alike.
Download (808kB)
![TOC [thumbnail of TOC]](http://repository.uph.edu/style/images/fileicons/text.png)
TOC.pdf
Restricted to Registered users only
Available under License Creative Commons Attribution Non-commercial Share Alike.
Download (708kB)
![Chapter 1 [thumbnail of Chapter 1]](http://repository.uph.edu/style/images/fileicons/text.png)
Chapter 1.pdf
Restricted to Registered users only
Available under License Creative Commons Attribution Non-commercial Share Alike.
Download (703kB)
![Chapter 2 [thumbnail of Chapter 2]](http://repository.uph.edu/style/images/fileicons/text.png)
Chapter 2.pdf
Restricted to Registered users only
Available under License Creative Commons Attribution Non-commercial Share Alike.
Download (983kB)
![Chapter 3 [thumbnail of Chapter 3]](http://repository.uph.edu/style/images/fileicons/text.png)
Chapter 3.pdf
Restricted to Registered users only
Available under License Creative Commons Attribution Non-commercial Share Alike.
Download (838kB)
![Chapter 4 [thumbnail of Chapter 4]](http://repository.uph.edu/style/images/fileicons/text.png)
Chapter 4.pdf
Restricted to Registered users only
Available under License Creative Commons Attribution Non-commercial Share Alike.
Download (2MB)
![Chapter 5 [thumbnail of Chapter 5]](http://repository.uph.edu/style/images/fileicons/text.png)
Chapter 5.pdf
Restricted to Registered users only
Available under License Creative Commons Attribution Non-commercial Share Alike.
Download (688kB)
![Bibliography [thumbnail of Bibliography]](http://repository.uph.edu/style/images/fileicons/text.png)
Bibliography.pdf
Restricted to Registered users only
Available under License Creative Commons Attribution Non-commercial Share Alike.
Download (688kB)
![Appendices [thumbnail of Appendices]](http://repository.uph.edu/style/images/fileicons/text.png)
Appendices.pdf
Restricted to Repository staff only
Available under License Creative Commons Attribution Non-commercial Share Alike.
Download (1MB)
Abstract
Perusahaan asuransi harus menentukan harga premi yang sesuai untuk setiap profil
risiko sehingga diperlukan model yang dapat memprediksi besar kerugian yang
mungkin terjadi di masa depan. Penelitian ini bertujuan untuk memproyeksikan
besar kerugian yang mungkin terjadi, membandingkan performa dari generalized
linear models (GLM) dengan distribusi Tweedie dan gradient tree-boosted
Tweedie model (TDboost), serta melakukan ratemaking dengan metode pure
premium. Terdapat data tahun 2015, 2016, 2017, dan 2018 dari sebuah perusahaan
asuransi kendaraan di Spanyol. Pada masing-masing tahun, akan dilakukan
pemodelan kerugian dengan GLM dan TDboost. Melalui analisis RMSE diperoleh
bahwa model TDboost memberikan hasil yang lebih baik dibandingkan GLM.
Melalui analisis indeks gini dan kurva Lorentz juga diperoleh bahwa hasil premi
murni dari model TDboost dapat menutupi ekspektasi kerugian dengan lebih baik.
Selain itu, melalui analisis faktor risiko diperoleh bahwa polis asuransi dengan
adanya pengendara kedua, kelompok pengendara berusia 18-29 tahun, dan
kelompok pengendara dengan pengalaman mengemudi 0-7 tahun memberikan
ekspektasi kerugian yang lebih besar daripada kategori yang lainnya. / Insurance companies must determine the appropriate premium price for each risk
profile, so a model that can predict the amount of losses that may occur in the
future is needed. This study aims to project the amount of losses that may occur,
compare the performance of generalized linear models (GLM) with the Tweedie
distribution and gradient tree-boosted Tweedie model (TDboost), and perform
ratemaking with the pure premium method. Data from 2015, 2016, 2017, and 2018
from a vehicle insurance company in Spain is used. In each year, loss modeling
will be carried out with GLM and TDboost. Through RMSE analysis, it is obtained
that the TDboost model provides better results than GLM. Through analysis of the
Gini index and Lorentz curve, it is also obtained that the pure premium results
from the TDboost model can cover loss expectations better. In addition, through
risk factor analysis, it was found that insurance policies with the presence of a
second driver, the driver group aged 18-29, and the driver group with 0-7 years of
driving experience provided greater loss expectations than the other categories.
Item Type: | Thesis (Bachelor) |
---|---|
Creators: | Creators NIM Email ORCID Kosasih, Michelle NIM01112210012 michellekosasih8@gmail.com UNSPECIFIED |
Contributors: | Contribution Contributors NIDN/NIDK Email Thesis advisor Saputra, Kie Van Ivanky NIDN0401038203 kie.saputra@uph.edu Thesis advisor Jobiliong, Eric NIDN0323067204 eric.jobiliong@uph.edu |
Uncontrolled Keywords: | generalized linear model; distribusi Tweedie; multi-year analysis; loss model; gradient tree-boosted Tweedie model; ratemaking. |
Subjects: | Q Science > QA Mathematics |
Divisions: | University Subject > Current > Faculty/School - UPH Karawaci > Faculty of Science and Technology > Mathematics Current > Faculty/School - UPH Karawaci > Faculty of Science and Technology > Mathematics |
Depositing User: | Stefanus Tanjung |
Date Deposited: | 09 Aug 2025 15:53 |
Last Modified: | 09 Aug 2025 15:53 |
URI: | http://repository.uph.edu/id/eprint/70435 |