路堑开挖爆破对民房危害的随机森林预测模型  

Random forest prediction model and its application to predicting house hazard from cutting excavation blasting

作　　者：李辉[1] 冯东梅[1] 马寒[1] 

机构地区：[1]辽宁工程技术大学工商管理学院,辽宁葫芦岛125105

出　　处：《辽宁工程技术大学学报（自然科学版）》2015年第12期1408-1413,共6页Journal of Liaoning Technical University (Natural Science)

摘　　要：为研究路堑开挖爆破对邻近民房安全的危害,运用主成分分析及随机森林算法对其进行预测.选取爆破参数、地质条件、民房结构3个方面的共16项重要影响因素,采用主成分分析法并从中提取6个主要成分.以主成分值为输入,房屋安全程度的量化值为输出,建立路堑开挖爆破对邻近民房安全危害的随机森林预测模型.利用18组工程实例数据为训练样本,另外4组数据为检验样本,进行了模型的预测实验.实验结果表明:基于主成分分析的随机森林模型对数据的拟合度较高,预测误差低,该模型可以作为实现路堑开挖爆破对邻近民房安全危害预测的一个有效方法.In order to do research on the harms to residential house safety brought by cutting excavation blasting, principal component analysis and random forest algorithm are applied to do the prediction. 16 important influence factors are selected from three aspects： blasting parameters, geological conditions, and the structure of the houses. Six major components are extracted from these factors by principal component analysis. Taking principal component values as input and quantitative values of safety degree as output, random forest prediction model is built to predict house safety hazard from cutting excavation blasting. Experiment is carried out using 18 sets of training sample and 4 sets of testing sample. The result shows that random forest prediction model based on principal component analysis has a high fitting degree and little error. The model is proved to be an efficient method for predicting residential safety hazard from cutting excavation blasting.

关 键 词：路堑开挖 爆破震动 民房安全 危害预测 主成分分析 随机森林 

分 类 号：X936[环境科学与工程—安全科学]