Comparative analysis of traffic flow prediction based on XGBoost before and after optimization

Due to the imbalance between the supply side and the demand side, the continuous urbanization process has challenged road traffic transportation, and improving the efficiency of road traffic transportation has become one of the challenges facing traffic engineering in the new era, in which the improvement of algorithmic prediction ability in machine learning has become one of the core of the solution. This study analyzes the accuracy and time benefit of XGBoost algorithm SSA optimization before and after two traditional deep learning in deep learning class models and simulates the fitting of time series feature data in the same space. From the simulation results, the models built in this study before and after optimization and the two traditional deep learning are within the error tolerance, and we believe that the prediction results using the optimized SSA-XGBoost are better for our dataset.