Application of a Combined Model based on K-means++ and XGBoost in Traffic Congestion Prediction

With the rapid increase in the number of vehicles, urban traffic congestion is becoming more and more severe. Therefore, it is particularly important to study road congestion and predict the changes of road congestion. A combined model based on K-means++ and XGBoost is proposed. In this method, average speed and traffic flow are used as the input of the model. The cluster center obtained by the K-means++ model and the value obtained by the XGBoost model are used to calculate the Euclidean distance. The minimum value of the calculation is used as the predictive value of the traffic congestion level. According to the prediction experiment of I15-N expressway traffic data in PeMS database, the combined model outperforms other models and the predictive accuracy of the combined model reaches 94.47%.