Sentiment-Analysis-with-scikit-learn-and-nltk

Uses voting from NaiveBayes, LinearSVC, MultinomialBinary, BernoulliNB, Logistic Regression for final result

About the training dataset

• The dataset is a collection of some ~10000 postive and negative reviews.

External Python Dependencies

• python7.7.7
• Scikit Learn@0.22.2
• nltk@3.5b1

Installing

• Run the jupter notebook Sentiment_Analysis_Notebook.ipynb. It will fill the pickled_files directory with saved models.
• Once done you can import sentiment_analysis.py module

Example Usage

• While inside the directory containing the sentiment_analysis.py file and the pickled_files folder

import sentiment_analysis as s

print(s.sentiment("The movies was very bad. They acting was horrible. Very bad experience! 0/10!"))