Survey on Soft Computing Methods for Accident Condition and Severity Predictions

Abstract

Roadways are a boon for both the agricultural and industrial sectors and are referred to as lifelines for any nation. The ever-increasing volume of transport and motor vehicles has literally choked these road networks resulting in traffic jams and road accidents. Road accidents are a combined result of the negligence of the vehicle driver, weather conditions, road condition and lighting conditions which results in a huge loss of life and property every year. Hence, there is a very strong need to develop an automated system to predict the severity of accidents and the possible reasons behind it. So, that preventive actions can be taken to avert these accidents in the future. This paper deals with the description of the prediction model and compares the model with the other soft computing methods to find out the most efficient method to predict the accident conditions and severity. SVM is found to have good average accuracy value among the available methods.