Category-Anchored Large Language Models for Next POI Recommendation
Read the original on arXiv AI →The paper introduces ANCHOR, a category‑anchored framework that improves next point‑of‑interest (POI) recommendation by first estimating category‑level mobility intent and then grounding this intent into a candidate set using category anchors, historical visit memory, and user preferences. A history‑aware reranker selects the final POI by matching candidates with the user’s past mobility patterns and current context. Experiments on three real‑world check‑in datasets demonstrate that ANCHOR consistently outperforms traditional, structure‑enhanced, and LLM‑based baselines, while offering favorable trade‑offs among candidate coverage, ranking quality, and inference efficiency.
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 AI.