NLTK POS tagger not working

Question

If I try this :

import nltk
text = nltk.word_tokenize("And now for something completely different")
nltk.pos_tag(text)

Output:

Traceback (most recent call last):
File "C:/Python27/pos.py", line 3, in <module>
nltk.pos_tag(text)
File "C:\Python27\lib\site-packages\nltk-2.0.4-py2.7.egg\nltk\tag\__init__.py" ipos_tag
tagger = load(_POS_TAGGER)
File "C:\Python27\lib\site-packages\nltk-2.0.4-py2.7.egg\nltk\data.py", line 605,in 
resource_val = pickle.load(_open(resource_url))
ImportError: No module named numpy.core.multiarray

score 10 · Accepted Answer · answered Jan 24 '13 at 18:22

10

It seems that the saved word tokenizer requires numpy. You'll need to install it.

answered Jan 24 '13 at 18:22

Danica

28,423
6
90
122

3

If Numpy's required by NLTK, shouldn't it be installed as a dependency automatically assuming one used Pip or other installer tools? I ask because I installed NLTK using Pip, then encountered the same issue. – Inactivist Feb 07 '13 at 15:51
1

@Inactivist Numpy's not required for most of NLTK, just some parts. (It's listed as optional on [the install page](http://nltk.org/install.html).) That said, it's super-handy to have around in general if you're doing any kind of machine learning type stuff or other number crunching. – Danica Feb 07 '13 at 16:09
wouldn't it be better for NLTK to split out the numpy-dependent code to a separate library – Jake Berger Jan 10 '15 at 20:04

score 3 · Answer 2 · edited Feb 24 '15 at 14:54

3

Install numpy using the command:

sudo pip install -U numpy

edited Feb 24 '15 at 14:54

kenorb

155,785
88
678
743

answered Feb 24 '15 at 14:45

Harshita Jhavar

147
8

NLTK POS tagger not working

2 Answers2

Linked

Related