Accident Severity Prediction in Big Data Using Auto-Machine Learning

Document Type : Article


1 Suleyman Demirel University, Graduate School of Natural and Applied Sciences, 32260, Isparta, Turkey

2 Suleyman Demirel University, Engineering Faculty, Department of Civil Engineering, 32260, Isparta, Turkey

3 Suleyman Demirel University, Property Protection and Security Department, 32260, Isparta, Turkey


Estimating the severity of a traffic accident is a problem in motor vehicle traffic because it affects saving human life. If the severity value can be predicted before the accident occurs, all the emergency teams needed could be sent to the area to provide faster first aid. With this aim, we studied a big data set for accidents in the USA between 2016 and 2020, which is almost 2.25x106 rows long. First, the data is preprocessed by removing the unnecessary variables. Then rows with blank cells are removed. Finally, about 1.7x106 rows length data is left for the prediction process. A machine learning algorithm has been used to determine the severity classification based on 16 input parameters. Moreover, binary to decimal count conversation has been used as a novel preprocessing method. As a result, the model has been built with a total accuracy of 0.816. the test results are also validated with precision, recall, and f1-score values. In this study, an auto-machine learning model has been developed and trained to predict the severity of a possible traffic accident based on the weather and road conditions.