e-vdb / movie-review-sentiment-classification

Sentiment analysis of movie reviews from Internet Movies Database.

movie-review-sentiment-classification

Summary

Sentiment analysis of movie reviews from Internet Movies Database (IMDb)

Dataset

Dataset from https://www.kaggle.com/lakshmi25npathi/imdb-dataset-of-50k-movie-reviews

50000 reviews : 25000 positive - 25000 negative

Streamlit interface

Repository content

• model/ folder containing different Python scripts
• moviereview-sentimentanalysis.ipynb: Jupyter Notebook

Notebook

• Natural Language Processing (NLP)
• Machine learning model : Logistic Regression
• Evaluation of the accuracy score

Visualization of the most discriminating features

Count Vectorizer (unigrams only)

Count Vectorizer (unigrams and bigrams)

Count Vectorizer with TFIDF rescaling (unigrams only)

Count Vectorizer with TFIDF rescaling (unigrams and bigrams)

TASK LIST

• Implement machine learning algorithm using Scikit-learn
• Implement deep learning algorithm using Keras
• Deploy model: Streamlit interface