Abstract
A popular research topic in Graph Convolutional Networks (GCNs) is to speedup the training time of the network. The main bottleneck in training GCN is the exponentially growing of computations. In Cluster-GCN based on this fact that each node and its neighbors are usually grouped ...
Read More
A popular research topic in Graph Convolutional Networks (GCNs) is to speedup the training time of the network. The main bottleneck in training GCN is the exponentially growing of computations. In Cluster-GCN based on this fact that each node and its neighbors are usually grouped in the same cluster, considers the clustering structure of the graph, and expand each node's neighborhood within each cluster when training GCN.The main assumption of Cluster-GCN is the weak relation between clusters; which is not correct at all graphs. Here we extend their approach by overlapped clustering, instead of crisp clustering which is used in Cluster-GCN. This is achieved by allowing the marginal nodes to contribute to training in more than one cluster. The evaluation of the proposed method is investigated through the experiments on several benchmark datasets.The experimental results show that the proposed method is more efficient than Cluster-GCN, in average.
Elena Touli; Oscar Lindberg
Abstract
In this paper, we relatively extend the definition of the global clustering coefficient to another clustering, which we call it \emph{relative clustering coefficient}. The idea of this definition is to ignore the edges in the network that the probability of having an edge is $0$. Here, we also consider ...
Read More
In this paper, we relatively extend the definition of the global clustering coefficient to another clustering, which we call it \emph{relative clustering coefficient}. The idea of this definition is to ignore the edges in the network that the probability of having an edge is $0$. Here, we also consider a model as an example that using the relative clustering coefficient is better than the global clustering coefficient for comparing networks and also checking the properties of the networks.
Mahmood Amintoosi
Abstract
A popular research topic in Graph Convolutional Networks (GCNs) is to speedup the training time of the network. The main bottleneck in training GCN is the exponentially growing of computations. In Cluster-GCN based on this fact that each node and its neighbors are usually grouped in the same ...
Read More
A popular research topic in Graph Convolutional Networks (GCNs) is to speedup the training time of the network. The main bottleneck in training GCN is the exponentially growing of computations. In Cluster-GCN based on this fact that each node and its neighbors are usually grouped in the same cluster, considers the clustering structure of the graph, and expand each node's neighborhood within each cluster when training GCN.The main assumption of Cluster-GCN is the weak relation between clusters; which is not correct at all graphs. Here we extend their approach by overlapped clustering, instead of crisp clustering which is used in Cluster-GCN. This is achieved by allowing the marginal nodes to contribute to training in more than one cluster. The evaluation of the proposed method is investigated through the experiments on several benchmark datasets.The experimental results show that the proposed method is more efficient than Cluster-GCN, in average.
