PERSONALIZED COMMENT RECOMMENDATION METHOD BASED ON GRAPH BIDIRECTIONAL AGGREGATION NETWORK LINK PREDICTION MODEL

Title:

PERSONALIZED COMMENT RECOMMENDATION METHOD BASED ON GRAPH BIDIRECTIONAL AGGREGATION NETWORK LINK PREDICTION MODEL

Document Type and Number:

WIPO Patent Application WO/2022/262190

Kind Code:

A1

Abstract:

A personalized comment recommendation method based on a graph bidirectional aggregation network link prediction model. Comment features are aggregated to user features in a user-comment bipartite graph. A social network is combined and user neighbor features are fused to obtain embedded representation of users. Then, the embedded representation of the users is aggregated into comments after removing original features of the users from the embedded representation of the users, and the embedded representation of the users is adjusted by comparing the difference value before and after the comments are aggregated. On this basis, a forwarding network is combined, the scores of edges are calculated by means of an inner product of user node features at the two ends of each edge, and finally recommendation is performed according to the scores. A recommendation system for assisting users is used to perform invisibility, and a comment recommendation task is converted into a link prediction task among the users in a small range. Moreover, a feature extraction method capable of processing different types of data is provided, such that a final result has higher representability.

More Like This:

WO/2024/020213	METHODS AND APPARATUS FOR NATURAL LANGUAGE-BASED SAFETY CASE DISCOVERY TO TRAIN A MACHINE LEARNING MODEL FOR A DRIVING SYSTEM
WO/2022/134794	METHOD AND APPARATUS FOR PROCESSING PUBLIC OPINIONS ABOUT NEWS EVENT, STORAGE MEDIUM, AND COMPUTER DEVICE
WO/2023/219765	SPLIT STAY RANKING AND MATCHING

Inventors:

SHU MINGLEI (CN)
WANG MUCHEN (CN)
WANG YINGLONG (CN)
LI ZHAO (CN)
GAO TIANLEI (CN)

Application Number:

PCT/CN2021/131595

Publication Date:

December 22, 2022

Filing Date:

November 19, 2021

Export Citation:

Click for automatic bibliography generation Help

Assignee:

SHANDONG ARTIFICIAL INTELLIGENCE INST (CN)
UNIV QILU TECHNOLOGY (CN)
SHANDONG COMPUTER SCIENCE CT NAT SUPERCOMPUTER CT JINAN (CN)

International Classes:

G06F16/9535

Foreign References:

CN113434756A	2021-09-24
CN103309972A	2013-09-18
CN103116639A	2013-05-22
US20070192350A1	2007-08-16
US20200160188A1	2020-05-21
CN112069402A	2020-12-11

Attorney, Agent or Firm:

JINAN SPRINGS PATENT&TRADEMARK OFFICE (CN)

Download PDF:

View/Download PDF PDF Help

Previous Patent: DRY ICE PELLET TRANSFER APPARATUS AND DRY ICE SLICE PRODUCTION DEVICE

Next Patent: DC FAULT CLEARING CONTROL METHOD AND DEVICE FOR HYBRID MMC, AND STORAGE MEDIUM