yelp-review-classifier

Introduction

This project uses scikit-learn to classify text data from Yelp reviews to predict what rating will be given. The application first preprocesses the Yelp data. Then the application creates a K-Folds cross validator with 20 splits. A Bagging Classifier is initialized with 5 Logistic Regression models.

A frontend is provided to visualize the results. A predicted star rating will display once at least 10 characters are typed.

Prerequisites

• scikit-learn
• Flask
• Pandas
• nltk

Usage

Clone this repository locally and train the model:

git clone https://github.com/kevintrankt/yelp-review-classifier.git
cd yelp-review-classifier
python api.py

Once the model is done training, visit yelp-review-classifier and start typing out a review to see the suggested rating!