基于随机森林的火灾损失影响因素研究

Research on factors affecting fire loss based on random forest algorithm

为使震后火灾损失评估的准确度更高,对火灾损失影响因素的重要度进行评估。从震后火灾的着火、蔓延和灭火3个方面提取地震烈度、房屋密度、人口密度、救火车密度、供水、反应速率和风速这7个影响因素,运用随机森林算法对其建模分析,得出了7个因素的重要度排序。反应速率和房屋密度最为重要。并依据重要度评估结果给出减少损失的措施。

In order to improve the accuracy of fire damage assessment after earthquakes,the importance of the factors affecting fire damage is evaluated.The seven influencing factors of seismic intensity,house density,population density,fire truck density,water supply,reaction rate and wind speed were extracted from the fire,spread and fire extinguishing of the post-earthquake.The random forest algorithm was used to model and analyze it.There are 7 factors ranked by importance.Among them,the reaction rate and house density are the most important.According to the results of the importance assessment,the corresponding measures to reduce losses are given.