Abstract:
Subgraph isomorphism problem is one of the most frequently encountered and extensively studied in the big graph database model and in Graph Theory. Graph Sub isomorphism problem is approached with machine learning techniques such as Graph Convolutional Network, representing the nodes in the form of embeddings such as node2vec and the Graph edit distance, for fi nding the dissimilarity of a Query node with respect to Target graph nodes. Using these a cost matrix is obtained and Munkres algorithm is applied to fi nd subgraph matching. Promising results are shown with these approaches, and deep learning techniques might be helpful to better approximate the cost matrix.
