Next-Generation Computing for Smarter Traffic: An Explainable AI Approach to Short-Term Flow Prediction

Accurate short-term traffic flow prediction is essential for optimizing transportation systems and mitigating urban congestion. Accurate short-term traffic flow prediction, which is critical for optimizing urban transportation, is evaluated using LSTM, GCN, and XGBoost models. To enhance interpretability, predictions were analyzed using Explainable Artificial Intelligence (XAI), specifically SHapley Additive exPlanations (SHAP). The study utilized the Traffic Flow Forecasting dataset from the UCI repository, which was preprocessed to ensure compatibility across models. Both performance and classification metrics were used to assess predictive accuracy. SHAP analysis provided valuable insights into the influence of various features and identified the key factors driving traffic flow predictions. The results highlight the importance of explainability in traffic prediction models, ensuring both transparency and reliability in intelligent transportation systems.