Work

Universal and non-universal text statistics: Clustering coefficient for language identification