Main Article Content
Public passenger transport holds immense significance in the overall transportation system. Forecasting the movement of public transport has emerged as a crucial problem in transport planning due to its practical implications. Recently, there has been a lot of significant attention in Intelligent Transportation Systems (ITS), introducing various advancements and innovative applications to develop conditions for public transit that are safer, more effective, and fun. To fully leverage the potential of ITS applications and deal with road situations proactively, it becomes crucial to have a reliable method for predicting traffic flow. This opens up opportunities for ITS applications to anticipate and address potential challenges in advance. Enhancing the efficient functioning of Public Transport (PT) networks is a primary objective for urban area authorities, and the proliferation of location and communication devices has led to an abundance of operational data. Applying appropriate Machine Learning (ML) methods can help identify patterns in the data to improve the Schedule Plan. This research focuses on heterogeneous information that influences the prediction value, aiming to predict the required transport demand for specific routes and the arrival time of public transport. Utilizing DBSCAN clustering with SARIMA Algorithm, real-time passenger demand forecasting is extensively promoted to enhance dynamic bus scheduling and management. Furthermore, this paper compares the accuracy of the proposed Prophet Model with traditional time series models like ARIMA and SARIMA. The aim is to provide precise and robust passenger demand predictions, enabling more effective planning and management of PT services.