Prediksi Tingkat Peringatan Kebakaran Semak di Perth Australia Barat menggunakan Metode Support Vector Machine dan Random Forest

Kanedi, Fidela Jovita (2025) Prediksi Tingkat Peringatan Kebakaran Semak di Perth Australia Barat menggunakan Metode Support Vector Machine dan Random Forest. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5026211112-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (4MB) | Request a copy

Abstract

Australia mengalami peningkatan cuaca ekstrem akibat suhu lebih tinggi dan kekeringan yang parah dalam beberapa dekade terakhir. Pada musim semi 2023, bagian barat daya Australia Barat mengalami suhu maksimum bulanan jauh di atas rata-rata untuk Agustus, September, dan Oktober, mencatat suhu terpanas sejak 1910 dan termasuk dalam 10% teratas tahun-tahun paling panas yang tercatat. Perth, ibu kota Australia Barat, adalah kota keempat terpadat di Australia pada 2020 dengan tingkat rawan kebakaran semak mencapai 90%, yang menjadi masalah serius karena populasi yang padat. Perubahan iklim meningkatkan frekuensi dan intensitas cuaca ekstrem seperti gelombang panas lebih panjang dan kekeringan lebih parah. Untuk meningkatkan respons terhadap kebakaran semak di Perth, pendekatan prediktif sangat penting. Prediksi yang akurat tentang potensi kebakaran membantu pihak berwenang mengalokasikan sumber daya secara efisien dan merencanakan strategi pemadaman yang tepat. Tugas Akhir ini bertujuan memprediksi tingkat peringatan kebakaran semak di Perth menggunakan metode Support Vector Machine (SVM) dan Random Forest. Data dikumpulkan melalui teknik data scraping dari laman resmi Biro Meteorologi Australia Barat, diproses, dianalisis, dan digunakan untuk melatih model. Model terbaik dari Random Forest dan SVM menggunakan proporsi data latih sebesar 80%, data uji sebesar 20%, dan sembilan fitur terbaik. Model Random Forest mampu menghasilkan akurasi 93,93% dan F1-score 93,77%, sedangkan Model SVM mampu menghasilkan akurasi 91,5% dan F1-score 91,5%. Model yang dikembangkan diharapkan membantu pemerintah meminimalisir kebakaran semak di masa depan.
=================================================================================================================================
Australia has experienced increased extreme weather due to higher temperatures and severe droughts in recent decades. In the spring of 2023, the southwest part of Western Australia saw monthly maximum temperatures well above average for August, September, and October, recording the hottest temperatures since 1910 and ranking among the top 10% of the hottest years on record. Perth, the capital of Western Australia, was the fourth most populous city in Australia in 2020 with a bushfire vulnerability rate reaching 90%, posing a serious issue due to its dense population. Climate change is amplifying the frequency and intensity of extreme weather events such as prolonged heatwaves and more severe droughts. To enhance the response to bushfires in Perth, predictive approaches are crucial. Accurate predictions of fire potential assist authorities in allocating resources efficiently and planning appropriate firefighting strategies. This research aims to predict bushfire alert levels in Perth using Support Vector Machine (SVM) and Random Forest methods. Data was collected through web scraping techniques from the official website of the Bureau of Meteorology Western Australia, processed, analyzed, and used to train the models. The best model of Random Forest and SVM used a proportion of training data of 80%, test data of 20%, and nine best features. The Random Forest model was able to produce an accuracy of 93.93% and an F1-score of 93.77%, while the SVM model was able to produce an accuracy of 91.5% and an F1-score of 91.5%. The developed model is expected to provide insights that can be utilized by relevant government authorities to reduce the incidence of bushfires.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Kebakaran Semak, Support Vector Machine, Random Forest, Cuaca, Confusion Matrix, Bushfire, Weather
Subjects:	T Technology > T Technology (General) > T57.8 Nonlinear programming. Support vector machine. Wavelets. Hidden Markov models.
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Information System > 57201-(S1) Undergraduate Thesis
Depositing User:	Fidela Jovita Kanedi
Date Deposited:	23 Jan 2025 01:21
Last Modified:	24 Jan 2025 04:15
URI:	http://repository.its.ac.id/id/eprint/116645

Actions (login required)

View Item