Recommendation method of local and global models based on feedback perception, terminal and medium

Abstract

The invention discloses a recommendation method for local and global models based on feedback perception, a terminal and a medium, and the method comprises the steps: capturing the local preference of a target object through a self-attention model of feedback perception; performing global preference learning according to the position attention layer and the local preference, and capturing the global preference of the target object; learning is carried out through a self-attention mechanism and a local intention model, modeling is carried out on the real intention of the target object, and a real intention model is obtained; and performing prediction according to the real intention model, the local preference and the global preference to obtain a next behavior prediction result of the target object. According to the method, different behaviors of the user are distinguished through the self-attention model of feedback perception, the local preference of the user can be captured, and the global preference of the user can be obtained through local preference learning; moreover, the corresponding article is matched based on the real intention, the real intention of the user in the next step is used as a query vector by utilizing local intention learning, and an accurate prediction result is obtained.

Description

technical field [0001] The present invention relates to the application field of data mining, in particular to a recommendation method, terminal and medium based on feedback-aware local and global models. Background technique [0002] In sequential single-class collaborative filtering, many isomorphic sequence recommendation algorithms have emerged, such as: RNN-based method GRU4Rec, CNN-based methods Caser and NextItNet, and attention-based algorithm SASRec. However, these methods cannot distinguish between different actions for the same item in a sequence because these algorithms are designed to only model a single type of action. [0003] To address this issue, some recent work on sequential heterogeneous single-class collaborative filtering algorithms attempts to model heterogeneous sequences, such as RLBL, RIB, and BINN. The cyclic log bilinear model (RLBL) divides a sequence into multiple time windows, uses the log bilinear model (LBL) to aggregate the interaction inf...

AI Technical Summary

Problems solved by technology

[0007] The technical problem to be solved by the present invention is that, aiming at the defects of the existing technology, the present invention provides a recommendation method, terminal and medium based on feedback-aware local and global models, so as to solve the technical problem that the existing heterogeneous prediction training is difficult

Images