US 12,229,681 B2
Trusted graph data node classification method, system, computer device and application
Yang Xiao, Xi'an (CN); Qingqi Pei, Xi'an (CN); and Zhuolin Xing, Xi'an (CN)
Assigned to XIDIAN UNIVERSITY, Xi'An (CN); and XI'AN XIDIAN BLOCKCHAIN TECHNOLOGY CO., LTD., Xi'An (CN)
Filed by Xidian University, Xi'an (CN); and Xi'an Xidian Blockchain Technology CO., Ltd., Xi'an (CN)
Filed on May 20, 2021, as Appl. No. 17/325,246.
Claims priority of application No. 202110028476.6 (CN), filed on Jan. 11, 2021.
Prior Publication US 2022/0222536 A1, Jul. 14, 2022
Int. Cl. G06N 3/084 (2023.01); G06N 3/047 (2023.01)
CPC G06N 3/084 (2013.01) [G06N 3/047 (2023.01)] 20 Claims
OG exemplary drawing
 
1. A trusted graph data node classification method, comprising:
inputting an adjacency matrix and a node feature matrix of a discrete topological graph, and calculating a discrete Ricci curvature of the discrete topological graph to extract topological information;
preprocessing of the discrete Ricci curvature and node features: preprocessing the discrete Ricci curvature and normalizing the node feature matrix;
mapping and normalizing the discrete Ricci curvature by using a multilayer perceptron (MLP) to obtain a mapped curvature matrix, reconstructing original features by using a feature reconstruction model to obtain a reconstructed feature matrix, performing a semi-supervised training by using the mapped curvature matrix and an original feature vector, and extracting and aggregating the node features; and
performing a classification prediction on nodes in the discrete topological graph by using a node classification model.
 
7. A computer device, comprising
a memory and a processor;
wherein
a computer program is stored in the memory, and when the computer program is executed by the processor, the processor executes the following steps:
inputting an adjacency matrix and a node feature matrix of a discrete topological graph, and calculating a discrete Ricci curvature of the discrete topological graph to extract topological information;
preprocessing of the discrete Ricci curvature and node features: preprocessing the discrete Ricci curvature and normalizing the node feature matrix;
mapping and normalizing the discrete Ricci curvature by using an MLP to obtain a mapped curvature matrix, reconstructing original features by using a feature reconstruction model, performing a semi-supervised training by using the mapped curvature matrix and an original feature vector, and extracting and aggregating the node features; and
performing a classification prediction on nodes in the discrete topological graph by using a node classification model.
 
8. A computer-readable storage medium, wherein
a computer program is stored in the computer-readable storage medium, and when the computer program is executed by a processor, the processor executes the following steps:
inputting an adjacency matrix and a node feature matrix of a discrete topological graph, and calculating a discrete Ricci curvature of the discrete topological graph to extract topological information;
preprocessing of the discrete Ricci curvature and node features: preprocessing the discrete Ricci curvature and normalizing the node feature matrix;
mapping and normalizing the discrete Ricci curvature by using an MLP to obtain a mapped curvature matrix, reconstructing original features by using a feature reconstruction model, performing a semi-supervised training by using the mapped curvature matrix and an original feature vector, and extracting and aggregating the node features; and
performing a classification prediction on nodes in the discrete topological graph by using a node classification model.