EVALUATION OF MACHINE LEANING ALGORITHMS FOR PREDICTING THE TRAFFIC CONGESTION

Abstract

Abstract: In this study, we have done the evaluation of the machine learning algorithms for predicting the traffic congestion. The XGBoost, LightGBM, Extra Tree, and Histogram Gradient Boosting machine learning algorithms were employed. The dataset was collected from the previous studies. On this collected data, feature engineering and K-mean clustering were applied to prepare it for the ML algorithms. The proposed method was developed in the Python language and used the Google Colab software for simulation purposes. The dataset was split into 70:30 for training and testing the ML algorithms. The simulation evaluation was done using error metrics, namely, Root Mean Square Error (RMSE), Mean Absolute Error (MAE), and determination of the coefficient (R²). The result indicates that XGBoost outperforms other ML algorithms employed in this study. Finally, the comparative analysis with the previous approach shows that the proposed method achieves a lower error metric value and a higher determination of the coefficient value due to preparing the dataset using feature engineering and k-mean clustering algorithms.