Enhancing Traffic Flow Prediction with Machine Learning Models: A Comparative Study

Precise traffic flow prediction is important for optimizing transportation management and developing urban strategies. This research study allows conducting a comparative analysis of various machine learning (ML) and deep learning (DL) algorithms to predict traffic flow patterns. The study examines five distinct models: Support Vector Regression (SVR), Extreme Gradient Boosting (XGBoost), Long Short-Term Memory (LSTM), Gated Recurrent Unit (GRU), and Convolutional Neural Networks (CNNs). SVR and XGBoost represent shallow and ensemble ML models, while the latter three exemplify DL architectures. The predictive quality of these models is evaluated utilizing four assessment metrics, Mean Squared Error (MSE), Root Mean Squared Error (RMSE), Mean Absolute Error (MAE), and coefficient of determination (<tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{R}^{\mathbf{2}}$</tex>). The analysis utilizes univariate traffic flow data sampled at 5 -minute intervals. The obtained results demonstrate the superiority prediction of the LSTM framework, achieving an <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{R}^{2}$</tex> value of 0.97. This comprehensive comparison provides valuable insights into the effectiveness of various prediction frameworks compressing machine and deep learning architectures for traffic flow prediction, thereby enabling transportation authorities and urban planners to make evidence-based decisions for improved urban mobility.