This is a flask web application that was created to utilize a machine learning sentiment classification model.
A data set consisting of 500,000 Amazon product reviews was used to fine tune a Multinomial Naive Bayes classifier.
Following training, the model had an accuracy rate of 93% on a test set. A K-means model was also created with 15 clusters to
analyze the data set with various visualizations.
To try the program out, visit https://flask-sentiment-app.herokuapp.com --> Enter test in the password field to login.