.png){kind=icon}
.png){kind=icon}
Return
Trajectory Classification Through Topological Data Analysis Perspectives
.png){kind=icon}
OA
AI
DOI：10.1109/ACCESS.2025.3543111.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
This paper examines the application of Topological Data Analysis (TDA) for trajectory classification, aiming to improve the interpretation of complex spatial movement patterns. By utilizing TDA, we explore the hidden structures in trajectory datasets, offering a fresh perspective on classification methods. Our study integrates TDA into trajectory analysis, highlighting its ability to capture spatial features that conventional methods may miss. We assess TDA's effectiveness using both simulated and real-world trajectory data from a survey comparing existing classifiers. TDA demonstrated significant performance improvements, with accuracy gains of up to 42.95% in certain scenarios. Notably, in real-world datasets, TDA increased accuracy by 38.49% for hurricane trajectory classification and improved precision by 39.24%. Simulated trajectories provided a controlled environment to further test TDA's robustness. The results underscore the potential of TDA to enhance trajectory analysis, uncovering complex spatial patterns and relationships that traditional methods may overlook.
Keywords:
Trajectory
Feature extraction
Software
Data analysis
Accuracy
Random forests
Long short term memory
Kernel
Computational complexity
Analytical models
Geometrical features
classification
clustering
trajectory analysis