A Rough Path Theory Approach to Classifying Geometric Paths

Read full text Summarise with VIRAT
Authors: Kangrong Tan
Cited by: 0

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

This study investigates the effectiveness of Rough Path Theory, specifically path signatures, for classifying geometric patterns, namely, distinguishing circular trajectories from linear or square trajectories. We generate synthetic two-dimensional path datasets consisting of 300 circular trajectories and 300 linear or square trajectories. Each path is embedded into a finite-dimensional feature space using its truncated signature, which captures both local and global geometric information. These signature features are subsequently employed to train a support vector machine (SVM) classifier with a linear kernel. The model achieves 100% classification accuracy on a 70/30 train-test split, with the confusion matrix indicating perfect separation between the two trajectory classes. The results confirm that path signatures encode sufficient information about curvature, ordering, and global topology to discriminate between different geometric trajectories. This demonstrates the potential of signature-based methods for shape recognition and time-series classification, where the geometry of observed paths, rather than raw coordinate values, carries essential diagnostic information.

Fields of Study & Subjects

Topological and Geometric Data Analysis Time Series Analysis and Forecasting Data Management and Algorithms

Publication details

Journal / source
Institutional Repositories DataBase (IRDB)
Volume
40
Pages
2–7
Publication date
1 March 2026
Language
EN