Postprint version. Published in Pattern Recognition Letters, Volume 24, Issue 1-3, January 1, 2003, pages 33-46. The definitive version can be found online at http://dx.doi.org/10.1016/S0167-8655(02)00186-1
The ‘LeRP’ algorithm approximates subgraph isomorphism for attributed graphs based on counts of length-r paths. The algorithm provides a good approximation to the maximal isomorphic subgraph. The basic approach of the LeRP algorithm differs fundamentally from other methods. When comparing structural similarity LeRP uses a neighborhood of nodes that varies in size dynamically. This approach provides sufficient evidence of similarity to permit LeRP to form a node-to-node mapping and can be computed with polynomial effort in the worst-case. Results from over 32,000 simulated cases are reported. We demonstrate that LeRP does not need a high dynamic range of node and edge coloring to perform well. For example, LeRP can input 50-node and 100-node graphs that contain a common 50-node subgraph, and then compute a matching subgraph having 49.74±0.46 nodes (mean ± one standard deviation). This takes from 0.4 to 0.5 s. In this example, 100 trials were evaluated and graphs had discrete coloring for nodes and edges with a dynamic range of four. Test conditions are varied and include strongly regular graphs as well as Model A.
Electrical and Computer Engineering