The purpose of this machine learning model is to predict the confidence in the policies that are being violated by the each customers review.
- Implemented in Python programming language
- Trained using RandomForest Regression algorithm.
- Used NLTK library for text cleaning and preprocessing.
- Achieved Root mean squared error of 0.01,0.03,0.001,0.03,0.01,0.01,0.01 and 0.001 on 8 multiple targets.
- Used grid search hyperparameter tuning technique to find best parameters for random forest
- Deployed as Web App using Django and Heroku.
- Used Bokeh Library to Analyse the confidence in the violated policies.
NOTE:The upload only accepts csv file and while uploading the file the column names should be city and review and there should not be any missing values in both the city as well as review column.
Link to Project : https://rishabh-second-test-app.herokuapp.com