Trajectory clustering based on FB-HDBSCAN algorithm

Aiming at the trajectory clustering problem of observed data, an improved Hierarchical Density-Based Spatial Clustering of Applications with Noise (HDBSCAN) algorithm, named FB-HDBSCAN, is proposed. Firstly, the spatiotemporal features of the data are thoroughly exploited to construct a distance matrix based on Fast Dynamic Time Warping (FastDTW), ensuring an accurate measurement of similarity between data points. Then, the Bayesian Optimization (BO) algorithm is introduced for adaptive parameter optimization, and the optimal parameters, along with the distance matrix, are input into the HDBSCAN algorithm framework to achieve automatic clustering of trajectory data. Experimental results demonstrate that the proposed algorithm achieves superior performance across multiple datasets, validating its feasibility and applicability.