Sentiment Classification using WSD, Maximum Entropy & Naive Bayes Classifiers

Hi,

I have used senti_classifier before and it worked fine. However, I had reinstalled it and since then it is giving zero score for anything and everything. Can anyone please help?

I'm trying to run:

from senti_classifier import senti_classifier
sentences = ['The movie was the worst movie', 'It was the worst acting by the actors']
pos_score, neg_score = senti_classifier.polarity_scores(sentences)
print pos_score, neg_score

but i keep getting error

Traceback (most recent call last):
  File "/home/robbie/sentiment/sentiment.py", line 1, in <module>
    from senti_classifier import senti_classifier
  File "build/bdist.linux-x86_64/egg/senti_classifier/senti_classifier.py", line 227, in <module>
  File "/usr/lib/python2.7/dist-packages/pkg_resources/__init__.py", line 1178, in resource_stream
    self, resource_name
  File "/usr/lib/python2.7/dist-packages/pkg_resources/__init__.py", line 1454, in get_resource_stream
    return io.BytesIO(self.get_resource_string(manager, resource_name))
  File "/usr/lib/python2.7/dist-packages/pkg_resources/__init__.py", line 1457, in get_resource_string
    return self._get(self._fn(self.module_path, resource_name))
  File "/usr/lib/python2.7/dist-packages/pkg_resources/__init__.py", line 1535, in _get
    return self.loader.get_data(path)
IOError: [Errno 0] Error: 'senti_classifier/data/SentiWn.p'

it looks like i don't have some sort of data file? Where do I get one or how do I make one? I thought it was already trained upon install.

Hi,

I am using sentiment_classifier in my project. I installed and importing in my script. but for all sentences it is getting 0,0 scores. I referred to last issue and verified if code version is latest and it is.

Can you help.

Attaching code and screenshot:

Code: pos_score, neg_score = senti_classifier.polarity_scores('The movie was the worst movie') print("Positive Score") print(pos_score) print("Negative Score") print(neg_score)

Screenshot:

pip install fails to add ".p" files on version 0.7 which leads to the error:

"FileNotFoundError: [Errno 2] No such file or directory: '...lib/python3.6/site-packages/senti_classifier/data/SentiWn.p'"

On downloading the latest build from https://pypi.python.org/pyp... and extracting, the "src/senti_classifier/data/" directory is missing the ".p" files. PyPi needs update.

Tried senti_classifier on some data to get sentiments of the sentences. It working fine for me. I still have two queries

1. It doesn't work for acronyms, Curious to know how can we add some acronyms in the dictionary of this algorithm.
2. Also, is there any way where it can handle negation?

Can you please let me know how can i overcome these? I am completely stuck at this point.

File "site-packages/pkg_resources/init.py", line 1576, in get_resource_stream return open(self._fn(self.module_path, resource_name), 'rb') FileNotFoundError: [Errno 2] No such file or directory: 'python3.5/site-packages/senti_classifier/data/SentiWn.p'

In addition to this, does polarity_scores() expect a paragraph string or a list of sentences? Because if it's the latter, there is more scope for optimization as an entire for loop can be removed.

Classifying text using the pickled classifiers is really slow. Like 6 seconds for a couple of sentences slow. Is it normal or am I doing something wrong? I'm following the README. Any tips to improve the performance?

Hi, I'm recently working on a project about sentiment analysis, and I would like to take this project as a reference. May I know where to input training data and how SentiWn.p is trained? Thanks a lot.

Hi,

If I try to import senti_classifier, I get the aforementioned error at line 229 - pickle.load(senti_pickle).

I have pulled latest code from master.

Regards, Chaitrali

I've cleaned up code a little bit for it to correspond PEP-8 best-practices, transformed it into python2/3 cross-compatible and wrote a simplistic utility for regression testing.

---------------------------------------------------------------------------
ModuleNotFoundError                       Traceback (most recent call last)
~\AppData\Local\Temp/ipykernel_29176/1367208207.py in <module>
----> 1 from senti_classifier import senti_classifier

C:\Project_Files\Anaconda3\lib\site-packages\senti_classifier\senti_classifier.py in <module>
    227 senti_pickle = resource_stream('senti_classifier', 'data/SentiWn.p')
    228 bag_of_words_pickle = resource_stream('senti_classifier', 'data/bag_of_words.p')
--> 229 synsets_scores = pickle.load(senti_pickle)
    230 bag_of_words = pickle.load(bag_of_words_pickle)
    231 bag_of_words = classify_polarity(bag_of_words)

ModuleNotFoundError: No module named 'collections\r'

I have tried pip install sentiment_classifier though pip install was successful, it didn't pull any files. So I referred to the Stack overflow post https://stackoverflow.com/questions/48521208/nltk-sentiment-classifier-issues-with-install and cloned GitHub repo and replaced the files into the senti_classifier folder. Now I get the above error .

Using Anaconda, Python 3.8, Windows 11