Analisis Risiko Kerugian Bencana Alam di Indonesia Menggunakan Extreme Value Theory-Value at Risk Dengan Pendekatan Bayesian Mixture Model

Hanin, Latifah Sekar (2026) Analisis Risiko Kerugian Bencana Alam di Indonesia Menggunakan Extreme Value Theory-Value at Risk Dengan Pendekatan Bayesian Mixture Model. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 2043221083-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (1MB) | Request a copy

Abstract

Indonesia merupakan salah satu negara dengan tingkat kerentanan bencana alam tertinggi di dunia, sehingga berpotensi menimbulkan kerugian ekonomi yang signifikan setiap tahun. Kondisi geologis dan geografis yang kompleks menyebabkan tingginya frekuensi bencana geofisika dan hidrologi, sehingga diperlukan analisis risiko yang mampu menggambarkan peluang terjadinya bencana sekaligus besarnya kerugian yang ditimbulkan. Penelitian ini menerapkan pendekatan kuantitatif melalui matriks risiko dengan dua komponen utama, yaitu frekuensi dan konsekuensi. Frekuensi kejadian bencana diestimasi menggunakan distribusi Poisson, yang sesuai untuk memodelkan kejadian yang bersifat jarang (rare events). Konsekuensi risiko diukur menggunakan VaR, yang merepresentasikan potensi kerugian pada tingkat kepercayaan tertentu. Karena data kerugian bencana bersifat ekstrem dan memiliki ekor distribusi yang berat (heavy-tailed), pemodelan VaR dilakukan menggunakan EVT dengan metode POT, di mana central part dimodelkan menggunakan distribusi Lognormal dan tail part menggunakan GPD. Estimasi parameter dilakukan dengan Bayesian Mixture Model melalui algoritma Metropolis–Hastings MCMC, dan parameter posterior digunakan dalam simulasi Monte Carlo untuk memperoleh estimasi EAL serta VaR. Hasil penelitian menunjukkan bahwa banjir dan gempa bumi berada dalam kategori High Risk. Kerugian tahunan rata-rata (EAL) akibat banjir mencapai 331.773 ribu USD, sedangkan gempa bumi mencapai 888.907 ribu USD, menunjukkan bahwa gempa bumi menimbulkan kerugian rata-rata yang lebih besar. Risiko kerugian ekstrem juga lebih tinggi pada gempa bumi, dengan VaR 95% sebesar 1.282.478 ribu USD untuk banjir dan 3.064.858 ribu USD untuk gempa bumi. Temuan ini menegaskan perlunya strategi mitigasi yang berkelanjutan serta alokasi dana kontinjensi yang lebih optimal untuk mengurangi potensi kerugian finansial akibat bencana alam di Indonesia.
=============================================================================================================================
Indonesia is one of the countries with the highest vulnerability to natural disasters, resulting in significant economic losses each year. Its complex geological and geographical conditions contribute to the high frequency of geophysical and hydrological disasters, making it essential to conduct risk analysis that captures both the likelihood of disaster occurrence and the magnitude of resulting losses. This study employs a quantitative approach using a risk matrix consisting of two main components: frequency and consequence. The frequency of disaster events is estimated using the Poisson distribution, which is suitable for modeling rare events. The consequence of risk is measured using VaR, which represents potential losses at a specified confidence level. Since disaster-related losses are extreme and heavy-tailed, VaR estimation is conducted using EVT with the POT method, where the central part of the data is modeled using the Lognormal distribution and the tail part using the GPD. Parameter estimation is performed through a Bayesian Mixture Model using the Metropolis–Hastings MCMC algorithm, and posterior parameters are utilized in Monte Carlo simulation to obtain estimates of EAL and VaR. The results indicate that floods and earthquakes fall into the High-Risk category. The average annual loss (EAL) from floods reaches 331.773 thousand USD, while earthquakes result in 888.907 thousand USD, indicating that earthquakes generate higher average losses. Extreme loss risk is also greater for earthquakes, with a 95% VaR of 1,282,478 thousand USD for floods and 3,064,858 thousand USD for earthquakes. These findings highlight the need for sustained disaster mitigation strategies and more effective allocation of contingency funds to reduce the financial impacts of natural disasters in Indonesia.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Bayesian, Bencana Alam, Extreme Value Theory, Mixture Model, Value at Risk, Bayesian, Natural Disasters, Extreme Value Theory, Mixture Model, Value at Risk
Subjects:	Q Science Q Science > Q Science (General) Q Science > Q Science (General) > Q180.55.M38 Mathematical models Q Science > QA Mathematics Q Science > QA Mathematics > QA279.5 Bayesian statistical decision theory.
Divisions:	Faculty of Vocational > 49501-Business Statistics
Depositing User:	Latifah Sekar Hanin
Date Deposited:	20 Jan 2026 08:50
Last Modified:	20 Jan 2026 08:50
URI:	http://repository.its.ac.id/id/eprint/129858

Actions (login required)

View Item