Article recommendation method and system based on user intention in session

Abstract

The present invention provides an article recommendation method and system based on user intent in a session by encoding a current session with articles of a previous timing and a last article, thereby generating an accurate session representation, and then using it to perform initial article prediction as user intent. Afterwards, we design an intentionally directed neighbor detector to locate a correct neighbor session. Finally, the representation of the current session and the representation of the neighbor session are combined through a gating fusion layer, final article recommendation is generated, and the accuracy of article recommendation is effectively improved.

Description

technical field [0001] The invention belongs to the technical field of content recommendation, and in particular relates to a method and system for item recommendation based on user intentions in a session. Background technique [0002] The existing item recommendation based on session content mainly refers to generating item recommendation based on anonymous user session. In order to improve the accuracy of anonymous session-based item recommendation, researchers have utilized collaborative information from the current session's neighbor sessions for the item recommendation. [0003] However, in the prior art, researchers usually directly select some sessions that are relatively close to the current session as neighbor candidate sessions, and they all consider the user intention in the session, and only use cosine similarity to collect and identify neighbor sessions , so the most relevant neighbors cannot be identified to obtain an effective neighbor representation, which ...

AI Technical Summary

Problems solved by technology

[0004] In view of this, the present invention provides a method and system for item recommendation based on user intentions in a s

Images