基于改进粒子群优化ν支持向量机的泥石流灾害预测模型  

Debris flow disaster prediction model based on improved particle swarm optimization ν support vector machine

作　　者：徐根祺 曹宁 李璐 谢国坤[1] 姚怡 Xu Genqi;Cao Ning;Li Lu;Xie Guokun;Yao Yi(School of Mechanical and Electrical Engineering,Xi'an Traffic Enginering Institute,Xi'an 710300,China;Civil Engineering College,Xi'an Traffic Enginering Institute,Xi'an 710300,China;School of Technology,Xi'an Siyuan University,Xi'an 710038,China;School of Urban Construction,Xi'an Kedagaoxin University,Xian710109,China)

机构地区：[1]西安交通工程学院机械与电气工程学院,西安710300 [2]西安交通工程学院土木工程学院,西安710300 [3]西安思源学院理工学院,西安710038 [4]西安科技大学高新学院城市建设学院,西安710109

出　　处：《国外电子测量技术》2022年第9期73-81,共9页Foreign Electronic Measurement Technology

基　　金：国家自然科学基金(51578461);西安交通工程学院中青年基金(2022KY-48)项目资助。

摘　　要：为了解决研究中存在的泥石流影响因子数据敏感度不同造成预测准确度不高、数据呈现小样本特点导致模型拟合效果不佳、非线性模型参数寻优困难等问题,采用主成分分析算法剔除相关性较强的影响因子,结合ν支持向量机(support vector machines, SVM)建立泥石流灾害预测模型,利用改进粒子群算法(particle swarm optimization, PSO)对模型进行优化,最终建立TFPSO-ν-SVM泥石流灾害概率预测模型。通过仿真实验对比了传统SVM模型、标准PSO-ν-SVM模型以及TFPSO-ν-SVM模型的性能,结果表明,TFPSO-ν-SVM模型具有最高的预测精度和最短的训练时间。In order to solve the problems that exist in the research, such as the data sensitivity of debris flow influencing factors is not high, the prediction accuracy is not high, the data is characterized by small samples, which leads to poor model fitting effect, and the optimization of nonlinear model parameters is difficult. The prediction model of debris flow disaster is established by combining with ν support vector machine(SVM), and the improved particle swarm algorithm(PSO)is used to optimize the model, and finally the TFPSO-ν-SVM debris flow disaster probability prediction model is established. The performances of the traditional SVM model, the standard PSO-ν-SVM model and the TFPSO-ν-SVM model are compared through simulation experiments. The results show that the TFPSO-ν-SVM model has the highest prediction accuracy and the shortest training time.

关 键 词：泥石流灾害 预测模型 粒子群算法 支持向量机 

分 类 号：P642[天文地球—工程地质学]