Naive bayes classifier for Youtube comments

Update July 2019: New link to app: https://am189.pythonanywhere.com/

Link to live web app on AWS Elastic beanstalk server:

http://mlspamapp.eu-west-2.elasticbeanstalk.com

This is a simple naive bayes classifier trained on structured data to detect spam in YouTube comments. The front-end is designed in HTML and CSS, while the back-end is written in the Python Flask web framework with Jinja2 templating.

Model accuracy: 91.95%

Screenshot of GUI:

Requirements: Python 3.7

Instructions:

Step 1:

Clone this project and set up a virtual environment https://docs.python.org/3/tutorial/venv.html

Step 2:

Install all relevant packages:

pip install flask

pip install flask_sqlalchemy

pip install pandas

pip install numpy

pip install scikit-learn

Step 3:

To run, open command line in folder and type:

python webapp.py

Go to localhost:5000 on your web browser to view and interact with app.

Things to do

• Deploy flask app to cloud server
• Add database functionality to store user input and retrain model - in progress