Data for this project is taken from Kaggle. It contains 50K Movie Reviews categorised into 'positive' and 'neagtive'. The review text is uncleaned and needs preprocessing. Find this data here.
The code is written in python using google colab. Find all the for the project here.
The analysis on this data is performed using Bernoulli Naive Bayes giving 87-88% accuracy and Multinomial Naive Bayes giving 88-89% accuracy.
The webapp to test this sentiment analysis model is built using streamlit. View streamlit docs here. Below is the image of the webapp desing.
