Estimasi Parameter Curah Hujan Ekstrem Di Jawa Barat Dengan Optimasi Bootstrap

Utami, Vania Dellafathi (2025) Estimasi Parameter Curah Hujan Ekstrem Di Jawa Barat Dengan Optimasi Bootstrap. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5006211075-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (17MB) | Request a copy

Abstract

Curah hujan ekstrem merupakan fenomena cuaca yang berpotensi menimbulkan bencana seperti banjir dan longsor, terutama di wilayah bertopografi kompleks seperti Jawa Barat. Untuk mendukung upaya mitigasi risiko, analisis dan prediksi curah hujan ekstrem dilakukan dengan menggunakan distribusi Generalized Extreme Value (GEV), yang mengestimasi parameter lokasi, skala, dan bentuk. Namun, ketidakpastian dalam estimasi parameter akibat variabilitas data dapat mengurangi akurasi model. Oleh karena itu, penelitian ini menerapkan metode bootstrap sebagai teknik resampling untuk meningkatkan stabilitas dan keandalan estimasi parameter distribusi GEV. Data curah hujan harian dari wilayah Bogor dan Sukabumi selama periode 2004–2024 dianalisis menggunakan metode block maxima per semester, dengan estimasi parameter dilakukan melalui pendekatan konvensional (Maximum Likelihood Estimation) dan bootstrap sebanyak 5000 iterasi. Hasil menunjukkan bahwa pendekatan bootstrap menghasilkan parameter yang lebih stabil dibandingkan metode konvensional, baik di Bogor maupun Sukabumi. Evaluasi performa model menggunakan Mean Absolute Percentage Error (MAPE) menunjukkan bahwa model bootstrap lebih mampu menangkap ketidakpastian dan variabilitas data namun tidak meningkatkan performa prediksi model, dengan nilai MAPE masing-masing sebesar 15,62% di Bogor dan 52,37% di Sukabumi, dibandingkan model konvensional yang mencapai 14,42% dan 51,18%.
========================================================================================================================================
Extreme rainfall is a weather phenomenon that has the potential to trigger natural disasters such as floods and landslides, especially in regions with complex topography like West Java. To support disaster risk mitigation efforts, the analysis and prediction of extreme rainfall are conducted using the Generalized Extreme Value (GEV) distribution, which estimates location, scale, and shape parameters. However, uncertainty in parameter estimation due to data variability can reduce the accuracy of the model. Therefore, this study applies the bootstrap method as a resampling technique to improve the stability and reliability of GEV parameter estimation. Daily rainfall data from the Bogor and Sukabumi regions during the 2004–2024 period were analyzed using the block maxima method on a semiannual basis, with parameter estimation performed through both conventional (Maximum Likelihood Estimation) and bootstrap approaches with 5,000 iterations. The results show that the bootstrap approach produces more stable parameter estimates compared to the conventional method in both Bogor and Sukabumi. Model performance evaluation using Mean Absolute Percentage Error (MAPE) indicates that the bootstrap model is better at capturing uncertainty and data variability but does not improve the model's predictive performance, with MAPE values of 15.62% in Bogor and 52.37% in Sukabumi, compared to 14.42% and 51.18% from the conventional model.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	curah hujan ekstrem, Generalized Extreme Value (GEV), bootstrap, mitigasi bencana, prediksi hidrologi,extreme rainfall, Generalized Extreme Value (GEV), bootstrap, disaster mitigation, hydrological prediction.
Subjects:	H Social Sciences > HA Statistics H Social Sciences > HA Statistics > HA31.7 Estimation
Divisions:	Faculty of Mathematics, Computation, and Data Science > Actuaria > 94203-(S1) Undergraduate Thesis
Depositing User:	Vania Dellafathi Utami
Date Deposited:	31 Jul 2025 02:19
Last Modified:	31 Jul 2025 02:19
URI:	http://repository.its.ac.id/id/eprint/124466

Actions (login required)

View Item