A Recommendation System Based on Sampling-Free Collaborative Knowledge Graph Network

Abstract

The present invention provides a recommendation system based on a non-sampling collaborative knowledge graph network, including: an embedding module is set to obtain the initial embedding vector of triples in the knowledge graph; the non-sampling knowledge graph convolution module is set to include several linear aggregators Single-layer convolutional network, pre-computing the initial embedding vector without sampling to obtain the deep-level information of the triplet; combining the embedding vector and the deep-level information as the updated embedding vector; the collaborative propagation module is set to encode user and item interactions simultaneously The collaboration signal in is used as the initial preference of users and items, combined with the updated embedding vector as the input vector of the prediction module; the prediction module is set to obtain the recommendation result according to the input vector. The present invention achieves a performance not inferior to the depth model, a faster speed, and a more accurate prediction result only by designing a relatively complex propagation matrix and a pre-computing operation.

Description

technical field [0001] The invention relates to the technical field of recommendation methods, in particular to a recommendation system based on a non-sampling collaborative knowledge graph network. Background technique [0002] The recommendation system is to use e-commerce websites to provide customers with product information and suggestions, to help users decide what products they should buy, and to simulate salespeople to help customers complete the purchase process. Personalized recommendation is to recommend information and products that users are interested in according to the user's interest characteristics and purchasing behavior. With the continuous expansion of e-commerce scale and the rapid growth of the number and types of products, customers need to spend a lot of time to find the products they want to buy. This process of browsing large amounts of irrelevant information and products will undoubtedly drain consumers who are drowning in information overload. ...

AI Technical Summary

Problems solved by technology

However, the introduction of GNN also faces the following problems: (1) The exponentially increasing number of nodes in the process of information dissemination leads to huge memory and time costs

However, the sampling operation may introduce errors in the optimization process

(2) Inherent problems such as gradient disappearance and feature smoothing in the deep graph neural network architecture make model training more difficult

Although some recent work has shown that these problems can be improved to some extent, extensive experiments have demonstrated that depth often does not bring significant benefits.

Images