Pushpendra9350 / Sentiment-prediction-on-reviews

This is a single webpage application in which we need to enter a review and this will tell you whether the review is positive or not. To make this work NLP techniques and Logistic regression algorithm is used with 94% accuracy

Sentiment prediction on reviews

Overview

This is a single webpage application in which we need to enter a review and this will tells you weather review is positive or not. To make this work NLP techniques and Logistic regression algorithm is used with 94% accuracy

Motivation

Motive to make this application

• To learn how to work with text data
• To learn NLP techniques
• To handle mid large dataset

To run this application

• First download the files specially model.pkl, app.py, and template folder.
• Just put them all in same directory
• Make sure you have installed all libraries written in requirements.txt file.
• On the terminal in the same folder run python app.py command then it will give you a url and go to that url now application is up and running.

Technical Aspects

Architecture

Future scope

• We can make histograms for most common questions to avoid reading all review for only one product.
• We can use this application to influence elections by knowing people sentimetal on twitter.
• we can use this application in any product