Abstract
Road traffic accidents are common in the world. India is among the leading countries in accident cases. Crash injury severity prediction is an exciting area of study in traffic safety. Traditional statistical models include underlying assumptions and preset relationships when it violated, might provide misleading findings. Due to many factors and negligence of traffic rules, many accidents are happening in India. This proposed work is aiming to determine the affecting factors/attributes in traffic accidents which provides a possible solution to the road traffic accident problem based on injury severity which in turns leads to the optimisation of the proposed model. Further, a feature selection method is applied to select the relevant road accident attributes for cognitive analysis. Feature selection increases the accuracy percentage of any classifier. In this research, ML (Machine Learning) and DL (Deep learning) algorithms are used for cognitive analysis and determining the factors which are mainly responsible for the happening of road traffic accidents in India. These algorithms include Support Vector Machine (SVM), Random Forest (RF), K Nearest Neighbours (KNN), Multi-Layer Perceptron (MLP) and Logistic Regression with 60%, 88%, 85%, 90% and 54% accuracy, respectively. The results have been evaluated using the various evaluation metrics such as Precision, Recall and F1-Score. Finally, the empirical result shows that the MLP algorithm of Deep Learning is better than the other proposed algorithms with an increased accuracy of approximate 88%.