摘要
车祸事故再现过程中,由于行人、风力等因素使得现场遗留痕迹中掺杂大量的与之不相干的数据信息,对车祸痕迹潜在挖掘信息形成了干扰。传统方法主要根据事故勘察报告和现场拍摄照片提取遗留痕迹信息进行事故再现,由于车祸事故的特殊性,不能有效地去除外部干扰数据信息,会导致车祸事故再现仿真真实度不高。提出考虑不确定性分析的车祸事故再现仿真挖掘方法。对车祸事故现场数据进行沃尔什离散化处理,并对其进行贝叶斯计算,针对获取的结果进行修正分析,完成车祸事故现场的特征提取。通过概率决策计算车祸事故现场数据关联度,计算不同车祸事故现场特征关联概率值,能有效去除现场遗留痕迹中干扰数据信息完成车祸事故再现仿真中挖掘结果不确定性分析。实验结果表明,利用改进算法进行车祸事故再现仿真中的挖掘结果的不确定性分析,能够提高车祸事故现场遗留痕迹挖掘准确性,提高了车祸事故再现仿真度。
A simulation mining method for car accident reappearance by considering the uncertainty analysis is proposed. The data of the car accident scene are made Walsh discretization and Bayesian calculation. The obtained results are made correction analysis to complete the extraction of the characteristics of the car accident scene. The de- gree of data correlation of car accident scene and the characteristics correlation probability value of the different car accident scene are calculated by probabilistic decision, which can effectively remove the interference data information of the remnant trace in the scene and complete the uncertainty analysis of the mining results of car accident reappear- ance simulation. Experimental results show that the improved algorithm can progress the accuracy of the remnant trace mining of car accident scene and increase the simulation degree of car accident reappearance.
出处
《计算机仿真》
CSCD
北大核心
2014年第10期187-190,共4页
Computer Simulation
关键词
车祸事故
关联规则
数据挖掘
遗留痕迹
Car accident
Association rules
Data mining
Remnant trace
