method for classifying graph structure data based on a network embedding algorithm and a CNN

A technology of embedding algorithm and network structure, applied in the direction of biological neural network model, calculation, neural architecture, etc., can solve the problems of research, lack of network classification, etc., and achieve the effect of easy training, good robustness, and good processing effect

Inactive Publication Date: 2019-03-26
BEIJING NORMAL UNIVERSITY +1
View PDF3 Cites 3 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, network embedding is more about the classification and prediction of nodes on the network, etc., and there is a lack of research on the classification of the entire network.

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • method for classifying graph structure data based on a network embedding algorithm and a CNN
  • method for classifying graph structure data based on a network embedding algorithm and a CNN

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0025] The technical details of this method are described in detail below.

[0026] The idea of ​​this method is to use the network embedding algorithm to convert the network representation into a high-dimensional space vector representation for the data classification problem of graph structure, and then convert it into a picture and then use the convolutional neural network CNN to realize the classification task. The process of this method is as follows figure 1 Next, take the international trade network classification as an example to introduce the specific steps in detail:

[0027] Step 1: Preprocessing of trade data

[0028] Our dataset is derived from the 4-digit code (SITC4) international trade dataset from 1962 to 2000 provided by the National Bureau of Economic Research (http: / / cid.econ.ucdavis.edu / nberus.html), covering The total trade volume of various products of various countries in 62-00, of which trade products can be divided into 10 categories, namely agricul...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention discloses a method for classifying graph structure data based on a network embedding algorithm and a CNN. Network embedding algorithm, graph structure data can be embedded into a high-dimensional space; Therefore, characteristics of the graph structure data are extracted, non-Euclidean data difficult to process are converted into Euclidean data easy to process, high-dimensional vector representation is processed into picture representation through a rasterization method, the picture representation is placed in a CNN to be trained and tested, and finally classification of the graph structure data is achieved. Compared with a traditional method, the method has the advantages that a deep learning method is applied, the classification accuracy is high, and more understanding is provided for the characteristics of the graph structure data of people. The method has great application in the aspects of social network, biological network, point cloud data processing and the like.

Description

technical field [0001] This scheme involves the classification of graph-structured data, and the study of computer algorithms for graph-structured data classification, specifically involving network embedding algorithms and CNN (Convolutional Neural Network) image classification techniques. Background technique [0002] Graph-structured data, that is, data composed of nodes and edges, represent objects and their relationships with each other. Today, graph-structured data is one of the most commonly used types of data. For example, relationships between people constitute social networks, road connections between cities constitute transportation networks, and citations between scientific papers constitute scientists Cooperation network, etc. [0003] The study of classifying graph-structured data is very interesting. First of all, we know that the current deep learning technology has achieved very good results in the processing of image data and text data. For example, CNN (...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
IPC IPC(8): G06K9/62G06N3/04
CPCG06N3/045G06F18/241G06F18/214
Inventor 辛茹月刘晶张江
Owner BEIJING NORMAL UNIVERSITY
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap