Suppose I have a directed graph G (V,E) whose transformation is defined by a library of patterns. Each vertex is of particular type.
The library of patterns contain subgraphs (g1,g2,g3 etc)and it's associated transformations(g'1,g'2,g'3 etc). I search the graph for each of these patterns and replace them with the associated transformed graph. After all detected patterns are substituted I get the resulting graph G'. Basically this is a search and substitute, of subgraphs.
The problem is this. I wish to use machine learning predict an estimate of the number of types of vertex in transformed graph. What could be possible features to extract ?