基于面板数据的交通事故宏观模型研究  

Traffic Accidents Macro Model Based on Panel Data

作　　者：王军雷[1] 孙小端[1] 贺玉龙[1] 徐婷[1] 

机构地区：[1]北京工业大学交通工程北京市重点实验室,北京100124

出　　处：《北京工业大学学报》2012年第4期575-580,共6页Journal of Beijing University of Technology

基　　金：国家自然科学基金资助项目(50778007)

摘　　要：作者基于7个国家1963—2005年的历史数据,分析了国内外交通事故的发展规律,阐述了宏观因素对交通安全的影响.通过模型比较,选择面板数据模型研究影响因素之间的关系,分别采用固定效应模型和随机效应模型进行估计.通过Husman检验表明,固定效应模型可以很好地描述交通事故十万人口死亡率与其他经济指标间的关系.结果表明,我国在当前经济发展GDP增长、汽车保有量增加的情形下,交通事故十万人口死亡率仍然会持续上升.To explore the relationship between traffic accidents and the macroeconomic factors, the authors collects historical data from seven counties from 1963 to 2005 to analyze the principle of traffic accidents. Panel data model is selected to measure the impacts of various factors after comparison because of its advantages of combinating time series models and cross section models. Fixed-effects model and random effects model are used to estimate accidents respectively. Husman tests are Used to decide which model is to choose. Results show that the fixed effect model well describes accidents and 100 000 population fatality rates, and other economic indicators. The discussion shows that as the economic development, China＇s current GDP and vehicles ownership keep growing in a few years and 100 000 population fatality rates continue to increase

关 键 词：面板数据 交通事故模型 宏观预测 固定效用 随机效应 Husman检验 

分 类 号：U491.31[交通运输工程—交通运输规划与管理]