Integrasi Pemodelan Sistem dengan Menggunakan Algoritma Kecerdasan Buatan untuk Mengidentifikasi Gempa Tektonik Multi-Event

Wati, Santi Ayu Rahma (2026) Integrasi Pemodelan Sistem dengan Menggunakan Algoritma Kecerdasan Buatan untuk Mengidentifikasi Gempa Tektonik Multi-Event. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 2042221089-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (3MB) | Request a copy

Abstract

Gempa bumi merupakan bencana alam yang sulit diprediksi karena kompleksitas data seismik serta hubungan spasial dan temporal antarkejadian yang bersifat dinamis dan tidak menentu. Metode konvensional memiliki keterbatasan dalam memodelkan pola nonlinier pada data gempa sehingga diperlukan pendekatan berbasis kecerdasan buatan. Penelitian ini bertujuan merancang sistem prediksi gempa tektonik secara *multi-event* menggunakan pendekatan *hybrid Artificial Intelligence* (AI) yang mengintegrasikan *Ant Colony Optimization* (ACO), *Genetic Algorithm* (GA), *Long Short-Term Memory* (LSTM), *Convolutional Neural Network* (CNN), dan *Naïve Bayes*. Data historis gempa bumi yang digunakan meliputi parameter magnitudo, kedalaman, waktu kejadian, serta koordinat lintang dan bujur yang diperoleh dari InaTEWS-BMKG. ACO digunakan untuk membentuk pusat aktivitas seismik dan area terdampak, sedangkan GA digunakan untuk menentukan arah dan sudut penyebaran gempa bumi. Seluruh hasil pemodelan kemudian diproses oleh LSTM sebagai model memori temporal untuk mengidentifikasi pola historis kejadian gempa serta mendeteksi anomali. Selanjutnya, CNN digunakan sebagai model utama untuk memprediksi arah dan sudut kejadian gempa berikutnya, sedangkan *Naïve Bayes* diterapkan untuk mengevaluasi keandalan hasil prediksi sistem. Hasil pengujian menunjukkan bahwa sistem yang dikembangkan mampu mengidentifikasi pola persebaran gempa bumi secara *multi-event* serta menghasilkan prediksi arah dan sudut berdasarkan data historis dengan tingkat keandalan yang baik. Integrasi berbagai algoritma kecerdasan buatan dalam satu sistem terbukti memiliki kemampuan yang memadai dalam memodelkan pola spasial dan temporal kejadian gempa bumi. Penelitian ini diharapkan menjadi landasan bagi pengembangan sistem prediksi gempa bumi berbasis data yang lebih adaptif guna mendukung upaya mitigasi bencana di masa mendatang.
================================================================================================================================
Earthquakes are natural disasters that are difficult to predict due to the complexity of seismic data and the dynamic, uncertain spatial and temporal relationships among events. Conventional methods have limitations in modeling nonlinear patterns in earthquake data; therefore, artificial intelligence-based approaches are required. This study aims to design a multi-event tectonic earthquake prediction system using a hybrid Artificial Intelligence (AI) approach that integrates Ant Colony Optimization (ACO), Genetic Algorithm (GA), Long Short-Term Memory (LSTM), Convolutional Neural Network (CNN), and Naïve Bayes. The historical earthquake data used in this study include magnitude, depth, occurrence time, and geographic coordinates (latitude and longitude) obtained from InaTEWS-BMKG. ACO is employed to identify seismic activity centers and affected areas, while GA is used to determine the direction and propagation angle of earthquake events. The modeling results are subsequently processed by the LSTM network as a temporal memory model to identify historical earthquake patterns and detect anomalies. CNN is then utilized as the primary model for predicting the direction and propagation angle of subsequent earthquake events, whereas Naïve Bayes is applied to evaluate the reliability of the prediction results. Experimental results demonstrate that the proposed system is capable of identifying multi-event earthquake distribution patterns and generating predictions of earthquake direction and propagation angle based on historical data with a satisfactory level of reliability. The integration of multiple artificial intelligence algorithms within a single framework demonstrates strong capability in modeling the spatial and temporal characteristics of earthquake occurrences. This study is expected to provide a foundation for developing more adaptive data-driven earthquake prediction systems to support future disaster mitigation efforts.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Prediksi Gempa, Kecerdasan Buatan, Mitigasi Bencana, Hybrid AI, Earthquake Prediction, Artificial Intelligence, Disaster Mitigation, Hybrid AI
Subjects:	H Social Sciences > HA Statistics > HA31.3 Regression. Correlation. Logistic regression analysis. Q Science Q Science > QA Mathematics > QA336 Artificial Intelligence Q Science > QA Mathematics > QA402.5 Genetic algorithms. Interior-point methods. Q Science > QA Mathematics > QA76.87 Neural networks (Computer Science)
Divisions:	Faculty of Vocational > Instrumentation Engineering
Depositing User:	Santi Ayu Rahma Wati
Date Deposited:	31 Jul 2026 09:26
Last Modified:	31 Jul 2026 09:26
URI:	http://repository.its.ac.id/id/eprint/140511

Actions (login required)

View Item