基于大数据的高速公路交通事故分析系统研究

Research on Expressway Traffic Accident Analysis System Based on Big Data

高速公路交通事故数据体量大、类型多、时效性要求高,应用大数据理论深度挖掘数据特征以提升运营安全已成为趋势。为建立交通事故分析系统,集合道路线形、交通流、历史事故数据等,首先采用邻近度技术、关联分析法清洗数据;然后从行车危险度角度,界定并筛选高速公路事故特征路段;最后利用变量柔性的广义负二项分布构建特征路段的事故预测模型。该系统的建立将可以充分挖掘事故数据特征,并辅助道路交通管理者实施决策。

Traffic accident data of expressway has large volume,many types,and high timeliness requirements.It has become a trend to apply big data theory to in-depth data features to improve operational safety.To establish a traffic accident analysis system,this paper collected road alignment,traffic flow,historical accident data,etc.Firstly,the data was cleaned using proximity technology and correlation analysis method.Then,from the perspective of driving risk,the highway accident characteristic road sections were defined and screened.Finally,the variable flexible generalized negative binomial distribution was used to construct the accident road prediction model of the characteristic road sections.The establishment of the system can fully excavate the characteristics of accident data and assist road traffic managers to implement decisions.