Analisis Survival Cox Proportional Hazard Dan Survival Support Vector Machine Pada Data Asuransi Jiwa Kredit Yang Mengalami Lapse Di PT Reasuransi X

., Ivan (2023) Analisis Survival Cox Proportional Hazard Dan Survival Support Vector Machine Pada Data Asuransi Jiwa Kredit Yang Mengalami Lapse Di PT Reasuransi X. Other thesis, Institut Teknologi Sepuluh Nopember.

[thumbnail of 06311940000053-Undergraduate_Thesis.pdf] Text
06311940000053-Undergraduate_Thesis.pdf - Accepted Version
Restricted to Repository staff only until 1 April 2025.

Download (2MB) | Request a copy

Abstract

Sektor perbankan di Indonesia memiliki peranan penting dalam pertumbuhan ekonomi dan pembangungan nasional serta kesejahteraan rakyat. Salah satu peran sektor perbankan dalam membangun perekonomian adalah melalui penyaluran kredit atau pinjaman. Penyaluran kredit sendiri tidak lepas dari risiko. Risiko yang mungkin terjadi adalah kredit macet. Salah satu penyebab peminjam tidak mampu membayar pinjaman mereka adalah kematian. Untuk memanajemen risiko tersebut, muncullah asuransi jiwa kredit. Asuransi jiwa kredit adalah produk kerja sama bank dengan perusahaan asuransi, yang memberikan manfaat berupa pelunasan kredit kepada bank apabila seorang yang memanfaatkan fasilitas kredit (debitur) meninggal dunia. Namun Asuransi jiwa kredit juga mungkin mengalami lapse yang mengakibatkan terhentinya perlindungan asuransi. Terjadinya lapse juga mengakibatkan hal buruk kepada perusahaan asuransi seperti berubahnya proyeksi cadangan dan mengurangi profit serta likuiditas. Pada penelitian ini, faktor penyebab lapse akan dianalisa menggunakan model Cox Proportional Hazard atau Cox Stratified dan metode survival support vector machine juga dilakukan untuk memperoleh prognostik indeks sehingga dapat dilakukan pengelompokkan observasi yang mengalami high risk dan low risk. Dengan menggunakan data asuransi jiwa kredit yang mengalami lapse di PT Reasuransi X dari periode 2017 – 2021, hasil permodelan Cox Proportional Hazard diperoleh bahwa uang pertanggungan, metode pembayaran, dan usia mempengaruhi kecendrungan terjadinya lapse dengan hazard ratio masing-masing sebesar 0,31589, 0,03368, dan 1,03080. Sementara itu pengklasifikasian dengan metode survival support vector machine berdasarkan median menghasilkan 1.149 observasi yang tergolong ke dalam kelompok high risk dan sisanya sebesar 1.148 data dikelompokkan menjadi kelompok low risk. Hasil pengelompokkan tersebut diuji kembali dengan menggunakan uji Log-Rank. Hasilnya terdapat perbedaan kurva survival antara kedua kelompok yang menunjukkan hasil pengelompokkan sudah baik. Metode survival support vector machine dengan kernel (Radial Basis Function) RBF menghasilkan nilai C-indeks terbaik yaitu sebesar 92%.
==============================================================================================================================
The banking sector in Indonesia plays a vital role in economic growth, national development, and the welfare of citizens. One of the roles of the banking sector in building the economy is through credit distribution or lending. Credit distribution itself is not free from risk, one of these possible risks being bad debt. A common cause of borrowers being unable to repay their loans is death. Due to this fact, credit life insurance emerges as a method of management for such risk. Credit life insurance is a product of cooperation between a bank and an insurance company, which provides benefits in the form of credit repayment to the bank should the credit facility user (the debtor) become deceased. However, credit life insurance may also experience a lapse which results in the cessation of insurance protection. The occurrence of lapses is also detrimental for the insurance companies, as it changes reserve projections, and reduces both profits and liquidities. This study attempts to analyze the factors causing the lapse using the Cox Proportional Hazard model. This model was chosen because the fulfillment of the distribution assumption at the time of survival is inessential, and only the fulfillment of the Proportional Hazard assumption is necessary. If these assumptions are not met, then the modeling will continue adopting the Cox Stratified model. The survival support vector machine method is also used to obtain a prognostic function, which allows the grouping of observations into high and low risk. By using credit life insurance data that has experienced a lapse at PT Reasuransi X from the period 2017 – 2021, the results of the Cox Proportional Hazard modeling show that sum assured, method of payment, and age affect the tendency for a lapse to occur with a hazard ratio of 0,31589, 0,03368, and 1,03080. Meanwhile, the classification using the survival support vector machine method based on the median resulted in 1,149 observations belonging to the high-risk group and the remaining 1,148 data grouped into the low-risk group. The grouping results were tested again using the Log-Rank test. The result is a difference in the survival curve between the two groups, which shows that the grouping results are promising. The survival support vector machine method with the (Radial Basis Function) RBF kernel produces the best C-index value of 92%.

Item Type: Thesis (Other)
Uncontrolled Keywords: Analisis Survival, Cox Proportional Hazard, Lapse, Survival Support Vector Machine, Survival Analysise
Subjects: Q Science > QA Mathematics > QA276 Mathematical statistics. Time-series analysis. Failure time data analysis. Survival analysis (Biometry)
Divisions: Faculty of Science and Data Analytics (SCIENTICS) > Actuaria > 94203-(S1) Undergraduate Thesis
Depositing User: Ivan .
Date Deposited: 18 Jan 2023 03:04
Last Modified: 18 Jan 2023 03:11
URI: http://repository.its.ac.id/id/eprint/95447

Actions (login required)

View Item View Item