Adaptive Trajectory Data Stream Clustering

Trajectory data mining is a field that focuses on analyzing and extracting insights from the movement patterns of objects over time. The realm of trajectory clustering algorithms has witnessed continuous evolution. Trajectory data streams have become integral in understanding movement patterns, but their real-time processing and clustering pose challenges. This research introduces a novel model for adaptive trajectory stream clustering with join rescheduling, batch monitoring, and simple pruning, which addresses these challenges for effective anomaly detection. Join rescheduling dynamically adjusts algorithm order, batch monitoring optimizes batch size, and simple pruning reduces noise. To validate the model, extensive experiments were conducted on real-world datasets movement, ge, and altitude. Ablation studies demonstrated the efficacy of each technique and revealed their combined strength in the proposed model. The results showcase enhanced clustering quality and improved resource utilization. The proposed model provides a comprehensive solution for efficient and accurate trajectory stream clustering.