36 lines
1.2 KiB
Plaintext
36 lines
1.2 KiB
Plaintext
|
|
=====================================================
|
|
PAICE's evaluation statistics for stemming algorithms
|
|
=====================================================
|
|
|
|
Given a list of words with their real lemmas and stems according to stemming algorithm under evaluation,
|
|
counts Understemming Index (UI), Overstemming Index (OI), Stemming Weight (SW) and Error-rate relative to truncation (ERRT).
|
|
|
|
>>> from nltk.metrics import Paice
|
|
|
|
|
|
-------------------------------------
|
|
Understemming and Overstemming values
|
|
-------------------------------------
|
|
|
|
>>> lemmas = {'kneel': ['kneel', 'knelt'],
|
|
... 'range': ['range', 'ranged'],
|
|
... 'ring': ['ring', 'rang', 'rung']}
|
|
>>> stems = {'kneel': ['kneel'],
|
|
... 'knelt': ['knelt'],
|
|
... 'rang': ['rang', 'range', 'ranged'],
|
|
... 'ring': ['ring'],
|
|
... 'rung': ['rung']}
|
|
>>> p = Paice(lemmas, stems)
|
|
>>> p.gumt, p.gdmt, p.gwmt, p.gdnt
|
|
(4.0, 5.0, 2.0, 16.0)
|
|
|
|
>>> p.ui, p.oi, p.sw
|
|
(0.8..., 0.125..., 0.15625...)
|
|
|
|
>>> p.errt
|
|
1.0
|
|
|
|
>>> [('{0:.3f}'.format(a), '{0:.3f}'.format(b)) for a, b in p.coords]
|
|
[('0.000', '1.000'), ('0.000', '0.375'), ('0.600', '0.125'), ('0.800', '0.125')]
|