Backward Feature Elimination

This workflow shows the implementation of the backward feature elimination procedure via the Backward Feature Elimnation metanode.

Inside the metanode you will find a loop extracting a subset of the input columns at each iteration according to the backeward feature elimination procedure.

The last node is the Backward Feature Elimination Filtering and allows to select the feature set and the corresponding accuracy.

External resources

• KDD Cup 2009 Data
• Seven Techniques for Data Dimensionality Reduction
• Random Forest for Data Dimensionality Reduction
• Paper Dissected: "Visualizing Data Using t-SNE" Explained
• Linear discriminant analysis
• Neural networks [6.1] : Autoencoder - definition
• Principal component analysis