ZURÜCK Word Frequencies in Language Zipfs Law

For a text corpus C let W(C) be the set of words occurring in C h(w) denote the frequency of the word wW(C) in the corpus. r(w) denote the rank of wW(C) if the words are ranked according to decreasing frequencies. It holds

r(w)·h(w)c = constant  wW(C)

ZUGANGAbb. 18: Zipfs Law applied to the Brown- and LOB-Korpus

ZUGANGAbb. 19: Qualitative View of Zipfs Law


© 1998 / HTML-Version 17. 11. 1998: R. Ferber