A study on traffic congestion prediction based on random forest model

Enhancing the precision and minimizing false alarms in predicting urban traffic congestion is imperative for maintaining a smooth traffic flow in cities. This paper explores the patterns of road congestion from both a spatial perspective and a temporal perspective. In the spatial perspective, this paper finds that there are differences in the traffic congestion values of different roads through the box plots of the traffic congestion values of each road, and then finds that there is spatial correlation between the traffic congestion values of different roads through the Pearson correlation analysis. In terms of time, this paper finds that the congestion data is cyclical. Thus, it is necessary to establish a model based on both temporal and spatial characteristics. In this paper, the Lasso model, Ridge model and random forest regression model are used to fit the traffic congestion data, and the model is evaluated by using three evaluation indexes, namely Mean Squared Error (MSE), Root Mean Squared Error (RMSE), and Mean Absolute Error (MAE), and the results of the study show that the random forest regression model is the most accurate in predicting the traffic congestion values.