Prediction of Forest Fire Occurrence in Peatlands using Machine Learning Approaches
Dedi Rosadi, Widyastuti Andriyani, Deasy Arisanty, Dina Agustina
Abstract
In this paper we consider the application of various machine learning approaches for prediction of the forest fire occurrence in the peatlands area. Here we consider some classical classification methods, such as support vector machine (SVM), k-Nearest Neighborhood (kNN), Logistic Regression (logreg), Decision Tree (DT) and Naïve Bayes (NB). For comparison purpose, we also consider more advanced algorithms, namely AdaBoost (DT based) approach. It is known that only a little number of similar studies is available for modeling peatlands fire occurrences in Indonesia. To illustrate the method, we consider the method using topographical and meteorological data from South Kalimantan Province. All computations are done using open source software R.