gnzeleven / Sentiment-Analysis-Naive-Bayes

Sentiment Analysis using Naive Bayes

Heroku web app

Using Naive Bayes for sentiment analysis on tweets. Given a tweet, the objective is to predict whether it has a positive sentiment or a negative sentiment.

Dataset: nltk's twitter_samples

• Train Naive Bayes model to analyse the sentiment of tweets
• Test the model's performance
• Predict on custom tweet

Install dependencies

Use the package manager pip to install all the dependencies.

pip install -r requirements.txt

Demo

Training and Testing

cd NaiveBayesSentiment
cd core/naivebayes

python naive_bayes_sentiment.py

Prediction

Click here

Deployment

• Credits
• Tutorial

Built With

• Django - Python based web framework that does all the heavy lifting
• Gunicorn - Python WSGI HTTP Server
• Heroku - Cloud application platform

Screenshots

Positive Example

Negative Example

Test Accuracy

ROC Curve

License

MIT