Model Classification of Fire Weather Index using the SVM-FF Method on Forest Fire in North Sumatra, Indonesia

As a tropical country, Indonesia is situated in Southeast Asia nation has vast forests. Forest fire occur busy vary due to land conditions and forest drought season. The indicator used mitigated potential study the behavior of weather index (FWI). data gathered from observation station north Sumatra province, computation estimation FWI by Canadian Fire Weather Index based on gathered. It found that there outlier data. hope will it, it necessary conduct classification predict this dataset machine learning approach using Support Vector Machine (SVM-FF), which further development previous models, known as c-SVM v-SVM. This method includes balancing parameter determining lower upper limits support vector. Furthermore, allowed value be negative. results showed was at low, medium, high, extreme levels. low an average 0.5 0 1 interval. There increase model’s accuracy performance its predecessor, include v-SVM with respective values 0.96 0.89. Meanwhile, observed SVM-FF model, quite better 0.99, indicating useful alternative classify fires.