Enhancing Traffic Accident Severity Prediction Using Artificial Intelligence Techniques

Every driver, irrespective of factors such as age, gender, driving experience and type of vehicle being used, faces a risk of being involved in traffic accidents. These accidents consist of incidents encompassing all types of vehicles such as cars, buses, motorcycles, bicycles, and trucks. and many times even pedestrians, resulting in about 1.35 million fatalities annually. Such accidents carry a noteworthy economic and social burden for the families of the victims. The accident severity factor plays a major role in incidents where deaths occur on the spot. Improvising the ability of predicting accident severity can benefit victims in getting a faster emergency response, thereby increasing their probability of surviving post impact. This paper analyzes the prediction methods of traffic accident severity using the Support Vector Machine (SVM) model using the classification learner application on MATLAB R2022b. Multiple factors were used in this analysis, namely age and gender of driver, types and numbers of vehicles involved in the accident, weather and street lighting conditions, day and time details. This model achieved an accuracy of 83.7%.