A decision tree algorithm for forest fire prediction based on wireless sensor networks. (10th August 2020)

• Record Type: Journal Article
• Authors: Gao, Demin; Xin, Jie; Zhang, Fuquan
• Is Part Of: International journal of embedded systems. Volume 13:Number 4(2020)
• Page Start: 422
• Page End: 430
• Publication Date: 2020-08-10
• Subjects: decision tree algorithm -- forest fire prediction -- wireless sensor networks -- WSNs; Embedded computer systems -- Periodicals; 004.16
• Languages: English
• ISSNs: 1741-1068

Abstract

Forest fire poses a significant threat not only to the natural environment and ecological systems but also to the safety of human life and property. Combined with new technologies, a decision tree algorithm is proposed for forest fire prediction, in which wireless sensor networks technology is utilised to transmit data and predict the ignition of the forest. There are four meteorological parameters as part of training data, containing temperature, relative humidity, wind speed, and daily precipitation, while the other part is prediction results of Forest Weather Index system. The decision tree generated by our system could classify these parameters from the most significant to the least significant so that it can better foretell fire occurrence. The analysis of prediction results shows that our system is effective.