Multi-class Multi-label Classification Model based on the ONET Job Database

less than 1 minute read

This project is a kind of real life use case: there is one database (ONET database) which is not designed to be used for machine learning, and we want to build a model that predicts a job title and abilities/skills from a job description. As the raw data is not initially framed for machine learning, we will have to extensively rearange it to get a dataset usable for machine learning. Then we will build a deep learning model using NLP and Keras. And finally we will use this model in a web application that we will deploy on HEROKU using Flask.

1. Dataset design and model creation

The Jupyter notebook below covers the dataset creation, and the model design.

ONET Job Database Model

2. Web application

The Web Application using the model parameters have been deployed on HEROKU, using Flask.

Web App