典型文献
Local feature aggregation algorithm based on graph convolutional network
文献摘要：
1 Introduction and main contributions
In the field of social networks and knowledge graphs,semi-supervised learning models based on graph convolutional networks have achieved great success in node classification[1],inductive node embedding[2],link prediction[3],and recommend.These semi-supervised models based on graph convolutional network(GCN)[4]expect to obtain more fea-ture information of a graph or accelerate the training.How-ever,we found that most of these models optimized the samp-ling method,neural network structure,and training mode,but paid no attention to the data preprocessing of the model.For example,the GraphSAGE[5]model aggregates adjacent objects within a certain distance from the central object to define the characteristics of the central object in training.Thus,the closer to the central object the neighbors are,the greater their role.
文献关键词：
中图分类号：
作者姓名：
Hao WANG;Liyan DONG;Minghui SUN
作者机构：
College of Computer Science and Technology,Jilin University,Changchun 130012,China;Key Laboratory of Symbolic Computation and Knowledge Engineering of Ministry of Education,Jilin University,Changchun 130012,China
文献出处：
引用格式：
[1]Hao WANG;Liyan DONG;Minghui SUN-.Local feature aggregation algorithm based on graph convolutional network)[J].计算机科学前沿,2022(03):197-199
A类：
samp
B类：
Local,feature,aggregation,algorithm,convolutional,Introduction,main,contributions,
In,field,social,networks,knowledge,graphs,semi,supervised,learning,models,have,achieved,success,node,classification,inductive,embedding,link,prediction,recommend,These,GCN,expect,obtain,more,information,accelerate,training,How,ever,we,found,that,most,these,optimized,ling,method,neural,structure,paid,attention,data,preprocessing,For,example,GraphSAGE,aggregates,adjacent,objects,within,certain,distance,from,central,define,characteristics,Thus,closer,neighbors,are,greater,their,role
AB值：
0.606945
相似文献
机标中图分类号，由域田数据科技根据网络公开资料自动分析生成，仅供学习研究参考。
