Traffic Mishap Injury Severity: An Unsupervised Approach

Predicting the injury severity of traffic accidents is of great significance in identifying the consequences on the victims of the accidents which may lead to physical or mental disturbance, financial losses, or the consequent health risks globally. Traffic mishaps are a number of the most essential troubles that the world is facing as they cause many deaths, injuries, and fatalities as nicely as financial losses each year. Identification of injuries severity is a key factor for the proper treatment as number of traffic accidents are increasing day by day and the victims are suffering from many major injuries even after many years of accidents. Unsupervised learning techniques are used in which association rule mining is applied on the Traffic Accident dataset which contains 660 records to generate frequent item sets. The strong rules contained in these regular item sets also show the relationship between influencing accident factors which can be used by breaking them to reduce the incidence of accidents. The rules can also be used to check the mishap scenes and some necessary safety measures can be made to avoid accidents and eventually improve the traffic safety of the city and providing assistance to doctors in treating patients. Further, a descriptive model is utilized for undertakings that would profit by the knowledge picked up from summing up information in new and fascinating manners. Therefore the Unsupervised learning techniques i.e Apriori Algorithm, Apriori- TID Algorithm, Eclat Algorithm revealed interesting patterns and knowledge that can be used in finding the association between accidents and related injuries.