A Decomposition-Driven Hybrid Framework Based on STL for Accurate Traffic Flow Forecasting

Accurate traffic flow forecasting is crucial for intelligent transportation systems and urban traffic management. However, existing single-model approaches often struggle to capture the complex, nonlinear, and multi-scale temporal patterns inherent in traffic flow data. This study proposes a novel decomposition-driven hybrid framework that integrates Seasonal-Trend decomposition using Loess (STL) with three complementary predictive models to enhance forecasting accuracy and robustness. The STL method first decomposes the original traffic flow time series into three distinct components: trend, seasonal, and residual. Subsequently, a Long Short-Term Memory (LSTM) network is employed to model the trend component and capture long-range temporal dependencies, an Autoregressive Integrated Moving Average (ARIMA) model is applied to the seasonal component to exploit periodic patterns, and an Extreme Gradient Boosting (XGBoost) algorithm is utilized to predict the residual component and handle nonlinear irregularities. The final forecast is obtained through multiplicative integration of the three sub-model predictions. The proposed framework is validated using 998 traffic flow records collected from a New York City intersection between November and December 2015. Experimental results demonstrate that the LSTM-ARIMA-XGBoost hybrid model significantly outperforms individual baseline models including standalone LSTM, ARIMA, XGBoost, and their variants (xLSTM, sLSTM, mLSTM) across multiple evaluation metrics, including Mean Absolute Error (MAE), Root Mean Square Error (RMSE), and coefficient of determination (R²). The decomposition strategy effectively isolates different temporal characteristics, enabling each specialized model to focus on its strength domain, thereby improving overall prediction accuracy, interpretability, and stability. This hybrid approach provides a reliable and efficient solution for real-time traffic flow forecasting in urban transportation systems.