Author : Mr. Abhijeet M. Tote Prof. Mr.Ranjit M. Gawande

The increase in road fatalities comes as bad news.  This cannot be stopped but can be controlled. The Accidents may cause due to driver’s health, Driver’s feelings, vehicle speed, climate condition, traffic conditions, road conditions, etc.  Analysis and prediction on Traffic collision has gained importance in these days. The big dataset is generated every year. The proposed system works on analysis and prediction of road accidents information data using machine learning algorithms and its efficient execution. For analysis same type of accidents are clustered together using EMM algorithm and for prediction association mining is performed using Improved Association Rule Mining (IARM) algorithm. The generated association rules are then provided to the Congestion control using Machine Framework (CCMF) and Traffic Congestion Analyzer using Map Reduce (TCAMP) algorithms to generate predictions. For efficient execution process, feature extraction technique and hadoop processing is used. The Results are compared in terms of accuracy and efficiency with existing systems.

Keywords: Road accidents, association rules ,big data ,hadoop, clustering, EMM, feature extraction