基于图嵌入与支持向量机的社交网络节点分类方法

Node classification method in social network based on graph embedding and support vector machine

针对无属性社交网络的节点分类问题,提出了一种基于图嵌入与支持向量机,利用社交网络中节点之间关系特征,对节点进行分类的方法。首先,通过DeepWalk、LINE等多种图嵌入模型挖掘节点隐含关系特征的同时,将高维的社交网络数据转换为低维embedding向量。其次,提取节点度、聚集系数、PageRank值等特征信息,组合构成节点的特征向量。然后,利用支持向量机构建节点分类预测模型对节点进行分类预测。最后,在三个公开的社交网络数据集上实验,与对比方法相比,提出的方法在社交网络节点分类任务中能取得更好的分类效果。

In order to solve the node classification problem for social networks without attributes,this paper proposed a method for classifying nodes based on graph embedding and support vector machine(SVM),which by using the relationship features between nodes in social network.Firstly,it obtained the implicit relationship features between nodes by using DeepWalk,LINE and other graph embedding models,and transformed the high-dimensional social network into low-dimensional embedding vector.Secondly,it extracted the node structure features such as degree,clustering coefficient,PageRank value and combined them to form the node feature vector.Thirdly,it used SVM to classify and predict the nodes.Finally,this paper conducted experiments on three real social network datasets,the results verify that the proposed algorithm improves the classification accuracy and has better classification effect.