Prediksi Risiko Kerugian Kebakaran Menggunakan Parametric Bootstrap

Abstract. Fire is a non-natural disaster that has a significant impact, both materially and non-materially. Extreme loss estimates are needed to support decision-making in risk mitigation and financial protection through insurance. This study aims to predict the maximum potential loss from a single extreme fire event using the Extreme Value Theory (EVT), Generalized Pareto Distribution (GPD), and parametric bootstrap approaches. The data used are fire insurance claims in Norway during the period 1972–1992, totaling 9181 incidents. The Peaks Over Threshold (POT) approach was used to identify extreme data above the threshold at the 90th percentile, which was 3602 (1000 Norwegian Kroner). The extreme data were modelled using GPD with parameter estimation via Maximum Likelihood Estimation (MLE), followed by parametric bootstrap to evaluate uncertainty and calculate Operational Value at Risk (OpVaR) at a 99% confidence level. The average OpVaR from 1000 iterations was 19920.33 (1000 Norwegian Kroner), with a 95% confidence interval between 18005.53 and 22197.63 (1000 Norwegian Kroner). The results indicate that the parametric bootstrap approach effectively enhances the reliability of the GPD model. This study contributes to the development of applied statistical methods in fire risk management and extreme event-based insurance strategy planning. Abstrak. Kebakaran merupakan bencana nonalam yang berdampak signifikan secara materi maupun nonmateri. Estimasi kerugian ekstrem diperlukan untuk mendukung pengambilan keputusan dalam mitigasi risiko dan perlindungan finansial melalui asuransi. Penelitian ini bertujuan memprediksi potensi kerugian maksimum dari satu kejadian kebakaran ekstrem menggunakan pendekatan Extreme Value Theory (EVT), Generalized Pareto Distribution (GPD), dan parametric bootstrap. Data yang digunakan adalah klaim asuransi kebakaran di Norwegia selama periode 1972–1992 sebanyak 9181 kejadian. Pendekatan Peaks Over Threshold (POT) digunakan untuk mengidentifikasi data ekstrem di atas ambang batas pada persentil ke-90, yaitu sebesar 3602 (1000 Krona Norwegia). Data ekstrem dimodelkan menggunakan GPD dengan estimasi parameter melalui Maximum Likelihood Estimation (MLE), dilanjutkan parametric bootstrap untuk mengevaluasi ketidakpastian dan menghitung Operational Value at Risk (OpVaR) pada tingkat kepercayaan 99%. Rata-rata OpVaR dari 1000 iterasi sebesar 19920.33 (1000 Krona Norwegia), dengan interval kepercayaan 95% antara 18005.53 hingga 22197.63 (1000 Krona Norwegia). Hasil menunjukkan bahwa pendekatan parametric bootstrap efektif memperkuat keandalan model GPD. Penelitian ini berkontribusi pada pengembangan metode statistik terapan dalam manajemen risiko kebakaran serta perencanaan strategi asuransi berbasis kejadian ekstrem.

Authors

Institutions

Publication Details

Journal
Bandung Conference Series Statistics
Published
2025-08-04
DOI
https://doi.org/10.29313/bcss.v5i2.20155
Primary Topic
Data Mining and Machine Learning Applications
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

Prediksi Risiko Kerugian Kebakaran Menggunakan Parametric Bootstrap

Tribuana Dewi Anugrah, Sutawanir Darwis
Bandung Conference Series Statistics
Data Mining and Machine Learning Applications
article

Prediksi Risiko Kerugian Kebakaran Menggunakan Parametric Bootstrap

Tribuana Dewi Anugrah, Sutawanir Darwis
article en

Abstract

Abstract. Fire is a non-natural disaster that has a significant impact, both materially and non-materially. Extreme loss estimates are needed to support decision-making in risk mitigation and financial protection through insurance. This study aims to predict the maximum potential loss from a single extreme fire event using the Extreme Value Theory (EVT), Generalized Pareto Distribution (GPD), and parametric bootstrap approaches. The data used are fire insurance claims in Norway during the period 1972–1992, totaling 9181 incidents. The Peaks Over Threshold (POT) approach was used to identify extreme data above the threshold at the 90th percentile, which was 3602 (1000 Norwegian Kroner). The extreme data were modelled using GPD with parameter estimation via Maximum Likelihood Estimation (MLE), followed by parametric bootstrap to evaluate uncertainty and calculate Operational Value at Risk (OpVaR) at a 99% confidence level. The average OpVaR from 1000 iterations was 19920.33 (1000 Norwegian Kroner), with a 95% confidence interval between 18005.53 and 22197.63 (1000 Norwegian Kroner). The results indicate that the parametric bootstrap approach effectively enhances the reliability of the GPD model. This study contributes to the development of applied statistical methods in fire risk management and extreme event-based insurance strategy planning. Abstrak. Kebakaran merupakan bencana nonalam yang berdampak signifikan secara materi maupun nonmateri. Estimasi kerugian ekstrem diperlukan untuk mendukung pengambilan keputusan dalam mitigasi risiko dan perlindungan finansial melalui asuransi. Penelitian ini bertujuan memprediksi potensi kerugian maksimum dari satu kejadian kebakaran ekstrem menggunakan pendekatan Extreme Value Theory (EVT), Generalized Pareto Distribution (GPD), dan parametric bootstrap. Data yang digunakan adalah klaim asuransi kebakaran di Norwegia selama periode 1972–1992 sebanyak 9181 kejadian. Pendekatan Peaks Over Threshold (POT) digunakan untuk mengidentifikasi data ekstrem di atas ambang batas pada persentil ke-90, yaitu sebesar 3602 (1000 Krona Norwegia). Data ekstrem dimodelkan menggunakan GPD dengan estimasi parameter melalui Maximum Likelihood Estimation (MLE), dilanjutkan parametric bootstrap untuk mengevaluasi ketidakpastian dan menghitung Operational Value at Risk (OpVaR) pada tingkat kepercayaan 99%. Rata-rata OpVaR dari 1000 iterasi sebesar 19920.33 (1000 Krona Norwegia), dengan interval kepercayaan 95% antara 18005.53 hingga 22197.63 (1000 Krona Norwegia). Hasil menunjukkan bahwa pendekatan parametric bootstrap efektif memperkuat keandalan model GPD. Penelitian ini berkontribusi pada pengembangan metode statistik terapan dalam manajemen risiko kebakaran serta perencanaan strategi asuransi berbasis kejadian ekstrem.

Bandung Conference Series StatisticsVol. 5(2)
Bandung Islamic University (ID)
Openalex Percentile: Top 68%
Data Mining and Machine Learning Applications
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.