Research of regional forest fire prediction method based on multivariate linear regression

Abstract

In order to achieve the predicted speed, high accuracy, the use of simple purpose, forest fire prediction of the key issues is to choose the main predictors. Forest fire prediction involves many factors, some of which are stable factors such as climate, topography, forest characteristics; and some unstable factors, such as fuel moisture content, meteorological factors, and other sources of ignition. Currently leading factor in the prediction of forest fire is often used in the fuel moisture, precipitation or dry days, relative humidity, temperature and wind five factors. In this paper, some of the data Yichun fire nearly a decade predict the forest fire meteorological data analysis, using multivariate linear regression to derive forest fire prediction method in the wireless sensor networks.