HowsMySpeech

A Sentiment Analysis Model Trained Using Sentiment140 Dataset. Currently The Model Is Only Capable Of Classifying Data Into Positive Or Negative Class. Followed A Lexical Based Approach To Train The Model, Achieving An Accuracy Of 82.28% On Test Data.

Requirements

Run pip3/pip install -r requirements.txt (Python 3)

Download The Sentiment140 Dataset From here

Upcoming Changes

1. Confidence Of Prediction Using Probability Prediction
2. Add Data To Train On Neutral Samples As Well.
3. Provide Web App Support To Check Analysis For A Twitter Hashtag Or Keyword.