Hi, Yes, if you use -1 as a placeholder for missing attributes, igraph considers that as a new type. How about simply taking the induced subgraph of all the vertices for which you know the gender and then just calculating the assortativity on that subgraph? Or, you could assign random genders to the vertices where you don't know the gender, take 1000 realizations and then take the average assortativity of these 1000 realizations.
> and one other question, in what condition do we get NaN ? For instance, if your graph has no edges, you will get a NaN. You will also get a NaN if all the types are identical. -- T. _______________________________________________ igraph-help mailing list [email protected] https://lists.nongnu.org/mailman/listinfo/igraph-help
