Abstract
Forest fire spreading is a complex burning phenomenon, and it is difficult to build a general spreading model for the fires occurred in different area over the world, even in the same country. Accordingly, predicting the burned area of forest fires is also a challenging task. In this work, five attributes (i.e. forest fuel moisture content, forest fuel inflammability, forest fuel load ,slope and burning time) are selected as input to predict burned area of forest fires occurred in the area of Guangzhou City in China. Next, using Data Mining (DM) technique, an SVM (Support Vector Machine) model was built and applied to deal with this type of a regression task, predicting burned area. Results showed that the selection of input attributes was reasonable, and the proposed SVM model was suitable for prediction of burned area, with higher precision, better generalization. This work provided a new way to deal with predictions for burned area of forest fires.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call