Traffic accident characteristics and association analysis of electric bicycles based on data mining

Abstract

With the increasing number of electric bicycles in cities, traffic safety is confronted with serious challenges. To prevent and control electric bicycle traffic accidents and further explore accident characteristics, this paper screens all 1555 general accident data records involving electric bicycles in Shenzhen from 2016-2021, 19 main impact factors are counted, and divided into three categories: accident information, personnel information, and road and facility information. Data mining is performed on the full accident set and each of the three single-dimensional accident sets: fatal accidents, escape accidents and accidents caused by electric bicycles. The Apriori algorithm is used to calculate and explore association rules, and the ones with better support, confidence and lift indexes are selected from them. From the association rules, this paper derives the relevant factors of electric bicycle traffic accidents, analyzes the coupling mechanism within the accidents, and provides suggestions on the countermeasures against the risk of electric bicycle traffic accidents.