Iterative Graph Embedding and Clustering
Graph embedding can be seen as a transformation of any graph into low-dimensional vector space, where each vertex of the graph has a one-to-one correspondence with a vector in that space. The latest study in this field shows a particular interest in a slightly different approach of graph embedding, where each node is inclined to preserve a community membership to respect high-order proximity and community awareness. We investigate different options of solving both tasks jointly, so a practical solution to one problem could be shared to enhance a solution to another problem and vice-versa. We imply that many iterations of such transferring can be made to achieve better results in both problems simultaneously. As a result of our work, we introduce a model that outperforms traditional methods which consider problems of graph embedding and community detection separately.