AGOUNI Takoua / SASSI Asma encadré par DJABRI Riadh2026-10-052026-06-08https://dspace.univ-tebessa.dz/handle/123456789/502Road traffic systems are essential in modern life, but the increasing number of vehicles has caused problems such as congestion and accidents. Therefore, traffic flow modelling is important for understanding and improving transportation systems. This thesis studies traffic flow modelling using input-output data and compares traditional methods with B-Spline Regression. Real traffic data collected from highways in France were used in the practical study. The results showed that B-Spline regression gives more accurate and flexible modelling of nonlinear traffic behavior, making it an effective tool for short-term traffic prediction.enTraffic FlowARXARMAXB-SplinesTraffic PredictionNonlinear ModellingThesis