基于SVM的驾驶员驾驶行为风险的研究  

A Study of Identification of Driver Behavior Risk Based on SVM

作　　者：余亚雄 陈玮[1] Yu Yaxiong;Chen Wei(College of Automation,Guangdong University of Technology,Guangzhou,Guangdong,510006,China)

机构地区：[1]广东工业大学自动化学院,广东广州510006

出　　处：《计算机科学与技术汇刊（中英文版）》2016年第1期46-50,共5页Transactions on Computer Science and Technology

摘　　要：近些年,随着车联网技术的发展,越来越多的车险产品开始基于车联网环境开发和应用,基于UBI（Usage Based Insurance）的车险便是其中的典型代表.作者提出将依赖于驾驶员驾驶风险评定的UBI车险费率厘定的问题转化为分类问题,并简单介绍了非线性支持向量机分类器的原理,然后针对驾驶员的风险评定问题,给出了构造的特征向量及解释,最后利用构造的特征向量对非线性支持向量机进行分类训练,利用训练的分类器辅助评定驾驶员的的出险概率,从而决定该驾驶员的UBI车险费用.Recently, with the development of telematics technological, the auto insurance based on the telematics technology is more andmore popular, the auto insurance based on UBI （Usage Based Insurance） is the typical of them. We propose that using the datamining technology to resolve the problem of the UBI auto insurance ratemaking, which relies on the identification of driverbehavior risk. In this paper, we firstly review the non-linear support vector machine （SVM） classifier, then we offer some featuresto train the SVM classifier, we could use the trained classifier to predict the probability of driver bahavior risk. Finally based onthese probability, we will be able to decide the UBI auto insurance premium for every driver.

关 键 词：UBI车险 数据挖掘 驾驶员行为风险 分类器 

分 类 号：TP[自动化与计算机技术]