TSMini: A Simple Yet Highly Effective Trajectory Similarity Learning Model

TSMini is a trajectory similarity learning model that introduces a sub-view modeling mechanism and a k‑nearest‑neighbor‑based loss to capture multi‑granularity trajectory patterns and relative similarity ranks. These innovations allow TSMini to approximate conventional trajectory similarity measures with high accuracy. Experiments demonstrate that TSMini outperforms state‑of‑the‑art models by an average of 15% on widely used similarity metrics.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv Machine Learning.

• benchmarks