abromberg/sentiment_analysis_python

Working with sentiment analysis in Python.

sentiment_analysis_python's Issues

Selecting features from all the data (both train and text data)

Hi,
I ran your program and found something that you may want to work on it ;)
when you are selecting the best features you should not look into your test data. It will look like your program is cheating :D
I did the same mistake once and i was very happy that my small program is beating all state of the art classification algorithms of the world.

Good Program though.

Coercing issue when the code is run

I am getting the following error , can you please help ?

---

TypeError Traceback (most recent call last)
in ()
60 #tries using all words as the feature selection mechanism
61 print 'using all words as features'
---> 62 evaluate_features(make_full_dict)
63
64 #scores words based on chi-squared test to show information gain (http://streamhacker.com/2010/06/16/text-classification-sentiment-analysis-eliminate-low-information-features/)

in evaluate_features(feature_select)
14 #http://stackoverflow.com/questions/367155/splitting-a-string-into-words-and-punctuation
15 #breaks up the sentences into lists of individual words (as selected by the input mechanism) and appends 'pos' or 'neg' after each list
---> 16 with open(RT_POLARITY_POS_FILE, 'r') as posSentences:
17 for i in posSentences:
18 posWords = re.findall(r"[\w']+|[.,!?;]", i.rstrip())

TypeError: coercing to Unicode: need string or buffer, RDD found