/spam-ham-web-app

A web app that classifies text as a spam or ham. I am using my own ML algorithm in the backend, Code to that can be found under machine_learning_section. For Live Demo: Checkout this link

Primary LanguageJupyter NotebookMIT LicenseMIT

SpamHam

SpamHam is a text-classification app which detects whether the message/email is spam or not. I've used Naive-Bayes along with NLP (TF-IDF, Bag of Words and more).
In order to perform an experiment I've combined two datasets (Enron email spam/ham and SMS spam classification) into one to gather more data. See this notebook to get what I am saying.
To check out this project in action I've deployed it on heroku Click on this link to check

Built With

  1. Django 2.1
  2. Python 3.6
  3. Scikit-Learn
  4. Numpy
  5. Pandas
  6. Matplotlib
  7. Seaborn
  8. HTML5
  9. CSS
  10. Bootstrap-v4
  11. Love

Installing/ Things you need to install the Web App and how to set up the project locally?

  1. Python3
  2. Pip
  3. Django(2.1)
  4. Conda

Steps

  • Make a virtual environment using "conda create -n envname python=3.6 pip"
  • source activate envname (for mac/linux) | activate envname (for windows)
  • Download or clone this repo by git clone https://github.com/aditya98ak/spam-ham-web-app.git
  • pip install -r requirements.txt
  • Run the app using python manage.py runserver

Milestones for version 2

  • Implement login and tailor experience for each user
  • Collect the result reported by user for false classification of messages/email
  • Model will self-learn from the reported data

Made with ❤️ by Aditya Kaushik - linkedin.com/adityakaushik001