A new proposal for graph classification using frequent geometric subgraphs

Geometric graph mining has been identified as a need in many applications. This technique detects recurrent patterns in data taking into account some geometric distortions. To meet this need, some graph miners have been developed for detecting frequent geometric subgraphs. However, there are few works that attend to actually apply this kind of pattern as feature for classification tasks. In this paper, a new geometric graph miner and a framework, for using frequent geometric subgraphs in classification, are proposed. Our solution was tested in the already reported AIDS database. The experimentation shows that our proposal gets better results than graph-based classification using non-geometric graph miners.