The Impact of Random Models on Clustering Similarity