Graph Compression Method Based on Feature Enhancement

Abstract

A graph compression method based on feature enhancement, which is applied to the classification of social network data sets. According to the end-to-end characteristics of the graph classification model, the method of calculating the gradient of the adjacency matrix of the input graph by the classification accuracy rate is used to obtain the model fitting function under The weight coefficients of all the edges are ranked according to the absolute value of the weight coefficients, and compared with the edge ranking calculated by the traditional edge importance index to calculate the overlap rate, the edge overlap rate is the largest Under certain circumstances, determine the number of key connections, keep the key connections, delete the remaining connections and isolated nodes to obtain a compressed graph, and then input the compressed graph into the same graph classification model for training and testing. Minimize the training time and computing space of the model as much as possible without reducing the classification accuracy much.

Description

technical field [0001] The invention relates to a graph compression method. Background technique [0002] Over the past few decades, research on graph-structured data has received increasing attention due to the collection of large amounts of structured data. In the study of graphs, a series of structural properties around nodes and edges have been proposed, including node centrality, clustering coefficient, associativity, similarity between pairs of nodes, etc. These properties are the basis of many graph-based The basis of the model. Furthermore, they capture certain local topological information of the system and thus can be used to design network algorithms. Generally, node centrality is always used to measure the importance of individuals in the system. Liben-Nowell and Kleinberg used the similarity measure of many nodes in social networks to predict whether there will be new interactions between them in the future (refer to literature 1: David Liben-Nowell, Kleinber...

AI Technical Summary

Problems solved by technology

[0004] At present, the graph classification model based on deep learning has achieved remarkable results. However, due to the huge amount of training data and the large number of nodes and edges in the graph, the running time and computing resources required for training are relatively high.

Images