Prediction of traffic accidents trend with learning methods: a case study for Batman, Turkey

Area of Science:

• Traffic Safety
• Transportation Engineering
• Data Science

Background:

• Road traffic accidents pose a significant public health and safety concern globally.
• Forecasting accident trends is crucial for effective prevention and control strategies.
• The province of Batman, Turkey, has not been previously studied for road accident trends.

Purpose of the Study:

• To analyze road traffic accident data in Batman, Turkey, from 2013 to 2022.
• To forecast road accidents, fatalities, and injuries for the next ten years (up to 2032).
• To evaluate the effectiveness of various machine learning models in predicting accident trends.

Main Methods:

• Analysis of road accident data including fatalities and injuries for drivers, passengers, and pedestrians.
• Application of machine learning models: State Space Models (SSM), Artificial Neural Networks (ANN), Autoregressive Integrated Moving Average (ARIMA).
• Utilized hybrid models: Convolutional Neural Networks (CNN) + Long Short-Term Memory (LSTM), and Attention + Gated Recurrent Unit (GRU).

Main Results:

• All models provided predictions with acceptable accuracy, though results varied based on parameters.
• The study indicates a continuous increase in road accidents, deaths, and injuries from 2022 onwards.
• Without intervention, accident numbers, fatalities, and injuries are projected to double by 2032.

Conclusions:

• Machine learning models can effectively forecast road accident trends.
• Urgent intervention is needed to address the rising road accident rates in Batman.
• Failure to mitigate causes of accidents could lead to a doubling of incidents by 2032.