Stacked Regression Model (XGBoost & Tabnet)

Inspiration

Inspired by existing ensemble models and a desire to create a new unique model that has the capability to make a greater prediction than the standalone XGBoost and Tabnet.

What it does

Machine learning model with a creative user interface in the form of a web page. Uses a metamodel (linear regression model) that takes in the outputs of the XGBoost and Tabnet predictions as input parameters, resulting in a robust ensemble model.

How I built it

Made in Python using machine learning libraries.

Built With

• anaconda
• jupyternotebook
• matplotlib
• numpy
• optuna
• os
• pathlib
• python
• pytorch
• shap
• sklearn
• streamlit
• tabnet
• vscode
• xgboost