Academy & Industry Research Collaboration Center (AIRCC)

Volume 10, Number 19, December 2020

Forest Fire Prediction in Northern Sumatera using Support Vector Machine Based on the Fire Weather Index


Darwis Robinson Manalu1, 2, Muhammad Zarlis1, Herman Mawengkang1 and Opim Salim Sitompul1, 1Universitas Sumatera Utara, Indonesia, 2Universitas Methodist Indonesia, Indonesia


Forest fires are a major environmental issue, creating economical and ecological damage while dangering human lives. The investigation and survey for forest fire had been done in Aek Godang, Northern Sumatera, Indonesia. There is 26 hotspot in 2017 close to Aek Godang, North Sumatera, Indonesia. In this study, we use a data mining approach to train and test the data of forest fire and the Fire Weather Index (FWI) from meteorological data. The aim of this study to predict the burned area and identify the forest fire in Aek Godang areas, North Sumatera. The result of this study indicated that Fire fighting and prevention activity may be one reason for the observed lack of correlation. The fact that this dataset exists indicates that there is already some effort going into fire prevention.


Forest fire, Fire Weather Index, Support Vector Machine, Machine Learning.