TRAFFIC SEVERITY PREDICTION USING MACHINE LEARNING AND DEEP LEARNING

Computer Science Engineering

Traffic accidents on highways remain a major cause of fatalities, even with advancements in traffic safety measures. The impact of injuries and damages from road incidents is especially severe in developing countries. Various factors lead to traffic accidents, with some significantly affecting the severity of these incidents. Data mining techniques can be instrumental in predicting the key factors linked to crash severity. This research pinpoints essential elements that correlate closely with accident severity on highways using Random Forest analysis. Key features influencing accident severity includeRange, heat level, cold breeze, moisture, clarity, and air movement.The study introduces a hybrid model that combines machine learning and deep learning methods, specifically Random Forest and Convolutional Neural Network, referred to as EFC(Ensemble Fusion Classifier) to forecast the severity of road accidents. The effectiveness of this model is evaluated against several baseline classifiers. The aim of this research is to improve the accuracy of predicting traffic accident severity by utilizing machine learning and deep learning techniques. The proposed EFC(Ensemble Fusion Classifier) model employs Random Forest for feature selection and a Convolutional Neural Network for enhanced pattern recognition, allowing for a thorough analysis of accident severity. This hybrid strategy enhances predictive capabilities by revealing complex relationships among contributing factors. The study emphasizes the benefits of merging machine learning and deep learning to create a dependable system for evaluating accident severity, which can assist traffic management authorities in implementing proactive safety measures.

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.