POI recommendation method and device based on hypergraph embedding
The invention discloses a POI recommendation method and device based on hypergraph embedding, and the method comprises the steps: receiving a historical track and a candidate POI list of a user, and mapping each node to a low-dimensional vector space through a hypergraph embedding method, so as to e...
Gespeichert in:
Hauptverfasser: | , , , , , , |
---|---|
Format: | Patent |
Sprache: | chi ; eng |
Schlagworte: | |
Online-Zugang: | Volltext bestellen |
Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
Zusammenfassung: | The invention discloses a POI recommendation method and device based on hypergraph embedding, and the method comprises the steps: receiving a historical track and a candidate POI list of a user, and mapping each node to a low-dimensional vector space through a hypergraph embedding method, so as to enable the similarity between different nodes to be calculated through the distance or cosine similarity between vectors; a space-time attention network is adopted to represent the relationship between the user behavior pattern and the POI, and the space-time attention network comprises two self-attention mechanisms and a bidirectional attention mechanism; and generating a recommendation result according to information learned by the space-time attention network. According to the method, the hypergraph embedding algorithm is used for regarding the sign-in information of the users as the hypergraph, random walk learning is carried out between the users to embed the vectors, the relevance between the sign-in informati |
---|