Authors : D. Kavitha, V. Kamakshi Prasad and J.V.R. Murthy
Abstract: In graph based data mining, graph/subgraph isomorphism testing used in mining frequent subgraphs plays key role and is time consuming. In a wide range of real applications, graph Isomorphism has significant role in retrieving the isomorphic graphs from a set of graphs. Canonical labelling of the graph has major impact on the efficiency of graph isomorphism testing. In this study, an algorithm is proposed to find canonical labelling in an efficient way and there by efficient isomorphism testing of labelled graphs. The proposed algorithm reduces search space based on the symmetries present in the graph there by making computation feasible to perform isomorphism testing on large databases for pattern mining.
D. Kavitha, V. Kamakshi Prasad and J.V.R. Murthy, 2016. Fast Graph Isomorphism Testing for Graph Based Data Mining with Improved Canonical Labelling. Journal of Engineering and Applied Sciences, 11: 1586-1597.