Pemodelan ratemaking asuransi kendaraan bermotor menggunakan metode regresi copula dan metode generalized linear model = Modeling motor vehicle insurance ratemaking using copula regression and generalized linear model

Buntaran, Carmen Ibanez Indrawati (2024) Pemodelan ratemaking asuransi kendaraan bermotor menggunakan metode regresi copula dan metode generalized linear model = Modeling motor vehicle insurance ratemaking using copula regression and generalized linear model. Bachelor thesis, Universitas Pelita Harapan.

	Text (Title) Title.pdf Restricted to Registered users only Available under License Creative Commons Attribution Non-commercial Share Alike. Download (249kB)
	Text (Abstract) Abstract.pdf Restricted to Registered users only Available under License Creative Commons Attribution Non-commercial Share Alike. Download (585kB)
	Text (ToC) ToC.pdf Restricted to Registered users only Available under License Creative Commons Attribution Non-commercial Share Alike. Download (615kB)
	Text (Chapter1) Chapter1.pdf Restricted to Registered users only Available under License Creative Commons Attribution Non-commercial Share Alike. Download (574kB)
	Text (Chapter2) Chapter2.pdf Restricted to Registered users only Available under License Creative Commons Attribution Non-commercial Share Alike. Download (685kB)
	Text (Chapter3) Chapter3.pdf Restricted to Registered users only Available under License Creative Commons Attribution Non-commercial Share Alike. Download (885kB)
	Text (Chapter4) Chapter4.pdf Restricted to Registered users only Available under License Creative Commons Attribution Non-commercial Share Alike. Download (1MB)
	Text (Chapter5) Chapter5.pdf Restricted to Registered users only Available under License Creative Commons Attribution Non-commercial Share Alike. Download (566kB)
	Text (Bibliography) Bibliography.pdf Restricted to Registered users only Available under License Creative Commons Attribution Non-commercial Share Alike. Download (566kB)
	Text (Appendices) Appendices.pdf Restricted to Repository staff only Available under License Creative Commons Attribution Non-commercial Share Alike. Download (4MB)

Abstract

Perusahaan asuransi perlu melakukan perhitungan premi yang harus dibayarkan oleh pemegang polis kepada perusahaan asuransi. Metode yang dapat digunakan untuk menghitung premi adalah dengan mengalikan nilai harapan klaim besar dengan nilai harapan frekuensi klaim, dengan asumsi independen antara besarnya dan frekuensi klaim. Penelitian ini menggunakan data polis asuransi kendaraan bermotor yang diamati antara Januari 2003 dan Desember 2004. Untuk setiap jenis klaim, dilakukan pemodelan menggunakan generalized linear model dan model regresi Gaussian copula untuk besar klaim dengan distribusi Gamma, Normal, dan Inverse Gaussian. Selain itu, frekuensi klaim dimodelkan menggunakan generalized linear model dengan distribusi Poisson. Berdasarkan analisis yang menggunakan root mean squared error (RMSE), secara umum, model generalized linear model dengan distribusi Inverse Gaussian-Poisson memberikan hasil RMSE yang paling kecil untuk premi murni dengan jenis klaim, yaitu TPL, damage, other, dan theft. Kemudian, model GLM Normal-Poisson memberikan hasil RMSE yang paling kecil untuk jenis jaminan fire, dan model regresi copula Inverse Gaussian-Beta-Normal dan Poisson memberikan nilai RMSE paling kecil untuk jenis jaminan windscreen. / Insurance companies need to calculate the premiums that must be paid by policyholders to the insurance company. The method that can be used is by multiplying the expected value of large claims by the expected value of claim frequency, assuming that the size and frequency are independent of claims. In this study, data from motor vehicle insurance policies between January 2003 and December 2004 were analyzed. Generalized Linear Model and Gaussian Regression models were used to model each type of claim. Copula was also used to model claim sizes with Gamma, Normal, and Inverse Gaussian. The frequency of claims was modeled using a Generalized Linear Model with a Poisson distribution. The results showed that the Generalized Linear Model with Inverse Gaussian-Poisson distribution provides the smallest root mean squared error (RMSE) results for pure premiums with claim types TPL, damage, other, and theft. For the fire collateral type, the Normal-Poisson GLM model gives the smallest RMSE results. On the other hand, the copula inverse Gaussian-Beta-Normal and Poisson regression models provide the smallest RMSE values for the windscreen collateral type.

Actions (login required)

View Item