Automatic learning of cost functions for graph edit distance

Abstract Graph matching and graph edit distance have become important tools in structural pattern recognition. The graph edit distance concept allows us to measure the structural similarity of attributed graphs in an error-tolerant way. The key idea is to model graph variations by structural distortion operations. As one of its main constraints, however, the edit distance requires the adequate definition of edit cost functions, which eventually determine which graphs are considered similar. In the past, these cost functions were usually defined in a manual fashion, which is highly prone to errors. The present paper proposes a method to automatically learn cost functions from a labeled sample set of graphs. To this end, we formulate the graph edit process in a stochastic context and perform a maximum likelihood parameter estimation of the distribution of edit operations. The underlying distortion model is learned using an Expectation Maximization algorithm. From this model we finally derive the desired cost functions. In a series of experiments we demonstrate the learning effect of the proposed method and provide a performance comparison to other models.