基于支持向量机回归多属性在地震应急决策中的应用探讨  

Application of Multi-attribute Regression Based on Support Vector Machine in Earthquake Emergency Decision Making

作　　者：王琳 杨天青[2] 邱鹏 郝婧 WANG Lin;YANG Tianqing;QIU Peng;HAO Jing(Guizhou Institute of Earthquake Engineering,Guiyang 550001,China;China Earthquake Networks Center,Beijing 100045,China)

机构地区：[1]贵州省工程防震研究院,贵阳550001 [2]中国地震台网中心,北京100045

出　　处：《华南地震》2022年第2期41-45,共5页South China Journal of Seismology

基　　金：国家重点研发计划(2018YFC1504506);六盘水市活动断层探测项目(黔科合支撑[2020]4Y053)联合资助。

摘　　要：地震应急是为了减轻地震灾害而采取的异于正常工作程序的紧急防灾和抢险行动。传统的震后应急决策主要依据应急条例标准以及决策者的经验,导致出队决策可能并非最优方案。随着数据挖掘和人工智能技术的发展,目前多种推荐算法在决策推荐领域正在被广泛的应用,对应急成员的基本情况按工龄、集合时间、应急次数3个属性进行分析,利用遗传算法全局最优收敛的特性,通过对支持向量机的3个属性参数进行寻优推荐,得到最优地震应急出队成员。Earthquake emergency is an emergency disaster prevention and rescue action which is different from normal working procedures to reduce earthquake disasters.The traditional post-earthquake emergency decisionmaking is mainly based on emergency regulations and the experience of decision makers,which may not be the optimal solution.With the development of data mining and artificial intelligence,various recommendation algorithms are widely used in the field of decision recommendation.In this paper,the basic situation of emergency response members is analyzed according to three attributes:length of service,gathering time and emergency response times.Using the characteristics of global optimal convergence of genetic algorithm,the optimal earthquake emergency team members are obtained by optimizing and recommending three attribute parameters of support vector machine.

关 键 词：地震应急 支持向量机回归 寻优推荐 

分 类 号：P511.2[天文地球—地质学]