Short-Term Traffic-Flow Forecasting Based on an Integrated Model Combining Bagging and Stacking Considering Weight Coefficient

This work proposed an integrated model combining bagging and stacking considering the weight coefficient for short-time traffic-flow prediction, which incorporates vacation peak time features, as well occupancy speed information, in order to improve prediction accuracy accomplish deeper traffic flow data feature mining. To address limitations of a single forecasting, with ridge regression meta-learner is first established, then optimized from perspective learner using model, lastly embedded into new base obtain Ba-Stacking model. Finally, model’s shortcomings terms low utilization, information structure learners modified by weighting error coefficients while taking account external resulting DW-Ba-Stacking that can change weights adjust distribution thus utilization. Using 76,896 I5NB highway empirical study object, compared assessed traditional this paper. The results show has highest accuracy, demonstrating successful predicting short-term flows effectively solve traffic-congestion problems.