Accident risk rating of streets using ensemble techniques of machine learning

Abstract

Increased vehicular traffic and lack of expert drivers on the street coupled with the adverse conditions and poor maintenance of streets are liable for increase in traffic accidents. Hence, prediction of traffic collision is of paramount importance for their mitigation. Street traffic analysis and prediction can be a dedicated approach to ensure safe and reliable street networks. The primary objective of this research is to assign an accurate accident risk factor for each street using machine learning models on the identified dataset. For automated and accurate prediction, various ensemble models of machine learning are applied, and their performance is compared with the naive models.