A highway congestion prediction method based on CNN-LightGBM

Most traditional highway congestion prediction methods mainly focus on constructing single prediction models and insufficiently explore spatiotemporal features. To address this problem, this paper proposes an ensemble model based on convolutional neural network (CNN) and light gradient boosting machine (LightGBM). Specifically, first, appropriate traffic congestion evaluation metrics are established to quantitatively analyze the level of road congestion; second, the CNN model is used to extract spatiotemporal features, which are then fed into the LightGBM model for prediction. To verify the effectiveness of the proposed model, various prediction models are compared in the experiments. The results demonstrate that the ensemble congestion prediction model achieves higher prediction accuracy. The proposed model is an efficient and fast approach for congestion prediction.