MBTI-predictor
A webapp using a linear SVM model to predict mbti personality types given a paragraph. The model was trained using MBTI-500 dataset. The model was 84% accurate on the test set.
Demo
This is how the main page looks like:
This is how the prediction page looks like:
Responsive design made to work on mobile as well
Running Locally
- Clone the repository:
git clone https://github.com/Proto007/MBTI-predictor/ - Create and activate a virtualenvironment: https://docs.python.org/3/library/venv.html#creating-virtual-environments
- Go to the directory of the project and install necessary dependencies:
cd MBTI-predictor
pip install -r requirements.txt
- Go to the django project directory and make database migrations:
cd mbti_predictor
python manage.py makemigrations
python manage.py migrate
- Run webapp in localhost:
python manage.py runserver
Tools used
Frontend: Django-template, Bootstrap CSS
Backend: Django-Rest-Framework
ML framework: scikit-learn