Prediksi Kebakaran Hutan dengan Analisis Deret Waktu Menggunakan Regresi Linier Berganda

Abstract

Forest fires are one of the most frequent natural disasters in Indonesia. These fires represent complex ecological disasters with multidimensional consequences. The resulting ecosystem damage can disrupt environmental balance and significantly impact social and economic aspects. Given the significant impacts of forest fires, early mitigation efforts are crucial. The application of regression algorithms in forest fire prediction can provide a scientific basis for more effective policy decision-making. One approach to implementing regression algorithms for forest fire prediction is through the use of multiple linear regression. This study aims to predict the burned forest area in a given year using a multiple linear regression model. The research is expected to contribute to the development of a more accurate early warning system for forest fires. This study was conducted through several stages, including literature review, data collection, data preprocessing, model implementation and evaluation, and conclusion and reporting. The targeted output is a web-based application to predict the burned forest area in Central Java. After optimization and testing scenarios, an average MAPE of 27% was obtained.