Statistics Based Predictive Geo-spatial Data Mining: Forest Fire Hazardous Area Mapping Application

Abstract

In this paper, we propose two statistics based predictive geo-spatial data mining methods and apply them to predict the forest fire hazardous area. The proposed prediction models used in geo-spatial data mining are likelihood ratio and conditional probability methods. In these approaches, the prediction models and estimation procedures depend on the basic quantitative relationships of geo-spatial data sets relevant to the forest fire with respect to the selected areas of previous forest fire ignition. In order to make the prediction map for the forest fire hazardous area prediction map using the two proposed prediction methods and evaluate the performance of prediction power, we applied a FHR (Forest Fire Hazard Rate) and a PRC (Prediction Rate Curve) respectively. When the prediction power of the two proposed prediction models is compared, the likelihood ratio method is more powerful than the conditional probability method. The proposed model for prediction of the forest fire hazardous area would be helpful to increase the efficiency of forest fire management such as prevention of forest fire occurrences and effective placement of forest fire monitoring equipment and manpower.