Difficult part: the most difficult part was easily the derivation of the generalized Zipf's law. In particular, I'm not entirely sure what the normalization condition mentioned in 3 is (though I have some idea).
Reflective part: the two papers seem to demonstrate, quite effectively, the perils of neglecting to thoroughly examine one's model to make sure it really supports what one wants to say. Zipf assumed that the correlation seen between word rank and frequency was due to some "law of economy", when in fact it derived inherently from the nature of his model.
Subscribe to:
Post Comments (Atom)
No comments:
Post a Comment