Using Principal Component Analysis and Decision Tree to Predict the Forest Fire Hazard Degree

摘要：

Prediction of the forest fire hazard degree is a key work for allocation rational of the forest fire prevention resource. A combination approach based on Principal Component Analysis(PCA) and decision tree is presented for predicting the forest fire hazard degree. The method of PCA was adopted to analysize the forest fire historical data of Guangzhou city. A new quantitative indexe system was selected by eliminating information overlap among variables and decreasing the index dimensions. Then choose 80 fire samples for the training of the decision tree to build the predictive model We used predictive model based on the rest 20 samples. The predictive accuracy is 95％. The results show that the model is effective to predict the forest fire hazard degree on Guangzhou city.

关键词： Principal Component Analysis decision tree2 forest fire3 hazard degree4 prediction

作者: Sun Yurong Zhang Gui

作者单位: College of Science,Central South University of Forestry and Technology,Changsha 410004,China

会议类型: 国际会议

会议名称: 2010 International Forum on Computer Science-Technology and Applications(2010 国际计算机科学技术应用论坛 IFCSTA 2010)

会议地点: 南宁

会议语种:英文

页码: 273-276

在线出版日期: 2010-12-10（万方平台首次上网日期，不代表论文的发表时间）