You can have a stop list; write a function to replace those words with space,or whatever you want; apply the function to the whole dataset.

Hi,

you might use your own stopwords file or nltk stopwords for example

stopwords=nltk.corpus.stopwords.words('english')

sentence=[ word for word in nltk.tokenize.wordpunct_tokenize(sentence.lower()) and word not in stopwords]

....

to add the step of tagging you might use python tagger or stanford tagger

see this for example:

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

tags_words=nltk.pos_tag(text)

words =[word[0] for word in tags_words if word[1] !='JJ' ]

>>> tags_words

[('And', 'CC'),

('now', 'RB'),

('for', 'IN'),

('something', 'NN'),

('completely', 'RB'),

('different', 'JJ')]

>>> words =[word[0] for word in tags_words if word[1] !='JJ' ]

>>> words

['And', 'now', 'for', 'something', 'completely']

http://www.nltk.org/book/ch05.html

Issa