PREDIKSI CURAH HUJAN DI KABUPATEN SIDOARJO MENGGUNAKAN METODE GA-LSTM DAN MITIGASI BENCANA BANJIR MENGGUNAKAN RISK MATRIX FLOOD ESTIMATION

Linggar, Dienna Eries (2025) PREDIKSI CURAH HUJAN DI KABUPATEN SIDOARJO MENGGUNAKAN METODE GA-LSTM DAN MITIGASI BENCANA BANJIR MENGGUNAKAN RISK MATRIX FLOOD ESTIMATION. Undergraduate thesis, UPN VETERAN JATIM.

[img] Text
cover skripsi.pdf

Download (3MB)
[img] Text
ISI-9-11 astrak.pdf
Restricted to Repository staff only until 28 July 2028.

Download (15kB)
[img] Text (BAB 1)
ISI-25-33 bab 1.pdf
Restricted to Registered users only until 28 July 2028.

Download (104kB)
[img] Text (BAB 2)
ISI-35-65 bab 2.pdf
Restricted to Repository staff only until 28 July 2028.

Download (578kB)
[img] Text (BAB 3)
ISI-67-83 bab 3.pdf
Restricted to Repository staff only until 28 July 2028.

Download (461kB)
[img] Text (BAB 4)
ISI-85-124 bab 4.pdf
Restricted to Repository staff only until 28 July 2028.

Download (940kB)
[img] Text (BAB 5)
ISI-125-127 bab 5.pdf

Download (80kB)
[img] Text (daftar pustaka)
ISI-129-133 dapus.pdf

Download (154kB)
[img] Text (lampiran)
ISI-134-135 lampiran.pdf
Restricted to Repository staff only until 28 July 2028.

Download (95kB)

Abstract

Climate change has become increasingly intense, leading to a rise in the frequency of extreme weather events, particularly heavy rainfall that can trigger flooding. Sidoarjo Regency, as a flood-prone area in East Java, requires a reliable daily rainfall prediction model to support risk-based disaster mitigation. This study develops and evaluates a daily rainfall prediction model using Long Short-Term Memory (LSTM) optimized with the Genetic Algorithm (GA), forming a hybrid GA LSTM model. GA is employed to determine the optimal combination of LSTM hyperparameters, such as the number of neurons, learning rate, batch size, and epochs, in order to enhance model performance. Daily rainfall data from 2020 to 2024 are used as the main dataset, with preprocessing stages including normalization and sequential data formation. The model is evaluated using Mean Squared Error (MSE), Root Mean Squared Error (RMSE), and Mean Absolute Error (MAE), along with k-fold cross-validation to ensure the stability and reliability of the predictions. The results show that the GA-LSTM model outperforms the standard LSTM, particularly in reducing squared error (MSE), with improved RMSE and MAE values. The predicted daily rainfall is then integrated with a Risk Matrix Flood Estimation to map flood risk levels both spatially and temporally in Sidoarjo Regency. This approach provides a robust foundation for more proactive, data-driven disaster mitigation decision-making. The study concludes that integrating GA-LSTM with a Risk Matrix can improve prediction accuracy and enhance flood mitigation effectiveness in tropical regions with high climate variability.

Item Type: Thesis (Undergraduate)
Contributors:
ContributionContributorsNIDN/NIDKEmail
AuthorMohammad Idhom, Mohammad IdhomNIDK19830310 202121 1 006idhom@upnjatim.ac.id
AuthorTrimono, TrimonoNIDK19950908 202203 1 003trimono.stat@upnjatim.ac.id
Subjects: T Technology > T Technology (General)
T Technology > T Technology (General) > T385 Computer Graphics
Divisions: Faculty of Computer Science
Depositing User: Dienna Linggar Eries
Date Deposited: 28 Jul 2025 08:02
Last Modified: 28 Jul 2025 08:02
URI: https://repository.upnjatim.ac.id/id/eprint/41311

Actions (login required)

View Item View Item