Linear Regression with 'autofeat' Engineered Features

This workflow demonstrates use of the component autofeat feature generator in feature generation. Problem pertains to predicting concrete strength. A Linear regression model is built using first the features already available and then using the features created by the autofeat generator. The model built using the generated features is much superior and can also be interpreted.

External resources

• Concrete Compressive Dataset at UCI
• Revisiting a Concrete Strength regression--Kaggle