PERBANDINGAN MODEL TWEEDIE DAN MODEL HURDLE DALAM ANALISIS FREKUENSI DAN BESARAN KLAIM ASURANSI KENDARAAN BERMOTOR

MAYA LESTARI PASARIBU, . (2026) PERBANDINGAN MODEL TWEEDIE DAN MODEL HURDLE DALAM ANALISIS FREKUENSI DAN BESARAN KLAIM ASURANSI KENDARAAN BERMOTOR. Sarjana thesis, UNIVERSITAS NEGERI JAKARTA.

	Text COVER.pdf Download (2MB)
	Text BAB 1.pdf Download (960kB)
	Text BAB II.pdf Restricted to Registered users only Download (212kB) | Request a copy
	Text BAB III.pdf Restricted to Registered users only Download (1MB) | Request a copy
	Text BAB IV.pdf Restricted to Registered users only Download (263kB) | Request a copy
	Text BAB V.pdf Restricted to Registered users only Download (545kB) | Request a copy
	Text DAFTAR PUSTAKA.pdf Download (515kB)
	Text LAMPIRAN.pdf Restricted to Registered users only Download (1MB) | Request a copy

Abstract

Penelitian ini bertujuan menganalisis frekuensi dan besaran klaim asuransi kendaraan bermotor menggunakan Model Tweedie dan Model Hurdle serta membandingkan kinerja kedua model berdasarkan nilai Akaike Information Criterion (AIC), Bayesian Information Criterion (BIC), Log-Likelihood, dan Root Mean Squared Error (RMSE). Data yang digunakan merupakan data sekunder klaim asuransi kendaraan bermotor sebanyak 678.013 observasi yang memuat variabel besaran klaim beserta beberapa variabel prediktor, yaitu lama pertanggungan, daya kendaraan, usia kendaraan, usia pengemudi, dan Bonus–Malus. Karakteristik data menunjukkan proporsi nilai nol yang tinggi, yaitu sebesar 94,98% pada frekuensi klaim dan 96,32% pada besaran klaim, serta distribusi besaran klaim yang menceng ke kanan (right skewed). Model Tweedie dibentuk menggunakan distribusi Compound Poisson–Gamma, sedangkan Model Hurdle terdiri atas komponen regresi logistik untuk memodelkan peluang terjadinya klaim dan komponen regresi Gamma untuk memodelkan besaran klaim positif. Estimasi parameter pada kedua model dilakukan menggunakan metode Maximum Likelihood Estimation (MLE). Hasil evaluasi menunjukkan bahwa Model Tweedie menghasilkan nilai AIC sebesar 645.320,20, BIC sebesar 645.411,60, Log-Likelihood sebesar −322.652,10, dan RMSE sebesar 5822,41. Sementara itu, Model Hurdle menghasilkan nilai AIC sebesar 637.138,90, BIC sebesar 637.264,30, Log-Likelihood sebesar −318.556,40, dan RMSE sebesar 5822,18. Berdasarkan hasil evaluasi tersebut, Model Hurdle memberikan kinerja yang lebih baik dibandingkan Model Tweedie karena memiliki nilai AIC dan BIC yang lebih kecil, nilai Log-Likelihood yang lebih besar (lebih mendekati nol), serta nilai RMSE yang lebih rendah. Kata kunci: Model Tweedie, Model Hurdle, frekuensi klaim, besaran klaim, asuransi kendaraan bermotor. ***** This study aims to analyze the frequency and severity of motor vehicle insurance claims using the Tweedie Model and the Hurdle Model and to compare the performance of both models based on the Akaike Information Criterion (AIC), Bayesian Information Criterion (BIC), Log-Likelihood, and Root Mean Squared Error (RMSE). The data used were secondary motor vehicle insurance claim data consisting of 678,013 observations, including claim severity and several predictor variables, namely exposure period, vehicle power, vehicle age, driver’s age, and Bonus–Malus. The data characteristics indicate a high proportion of zero values, namely 94.98% for claim frequency and 96.32% for claim severity, while the positive claim severity exhibits a right-skewed distribution. The Tweedie Model was constructed using the Compound Poisson–Gamma distribution, whereas the Hurdle Model consisted of a logistic regression component to model the probability of claim occurrence and a Gamma regression component to model positive claim severity. Parameter estimation for both models was performed using the Maximum Likelihood Estimation (MLE) method. The evaluation results showed that the Tweedie Model produced an AIC value of 645,320.20, a BIC value of 645,411.60, a Log-Likelihood value of −322,652.10, and an RMSE value of 5822.41. Meanwhile, the Hurdle Model produced an AIC value of 637,138.90, a BIC value of 637,264.30, a Log-Likelihood value of −318,556.40, and an RMSE value of 5822.18. Based on these evaluation results, the Hurdle Model outperformed the Tweedie Model, as it achieved lower AIC, BIC, and RMSE values, as well as a higher Log-Likelihood value (closer to zero), indicating better overall model performance in analyzing the frequency and severity of motor vehicle insurance claims. Keywords: Tweedie Model, Hurdle Model, claim frequency, claim severity, motor vehicle insurance.

Item Type:	Thesis (Sarjana)
Additional Information:	1). Ibnu Hadi, M.Si. ; 2). Dr. Yudi Mahatma, M.Si.
Subjects:	Sains > Matematika
Divisions:	FMIPA > S1 Matematika
Depositing User:	Maya Lestari Pasaribu .
Date Deposited:	31 Aug 2026 08:23
Last Modified:	31 Aug 2026 08:23
URI:	http://repository.unj.ac.id/id/eprint/73390

Actions (login required)

View Item