Model Stacking Workbench

Purpose

This is a collection of Python modules that facilitate building "a predicitve model" for Kaggle competitions using a generalized model stacking approach. This approach involves developing several predictive models and combining predictions from those models into an overall prediction. Generally speaking the combination of predictions lead to a better performing model.

This version of the workbench supports supports using model algorithms from scikit-learn and xgboost for regression and classification problems. Future releases will incorporate algorithms from other Python packages, such as TensorFlow, and R.

See project wiki for additional details.

Feedback will be appreciated. Feel free to enter comments in this repo's Issues tab.