neurodata / SPORF

This is the implementation of Sparse Projection Oblique Randomer Forest
https://neurodata.io/forests/
97 stars 46 forks source link

Bug in the demo "OpenML Datasets: Banknote example". #355

Open ZebinYang opened 3 years ago

ZebinYang commented 3 years ago

The target feature is mistakenly included in the training data.

It appears in cell [4] below "Returned as numpy array, with meta-info (e.g. target feature, feature names,…)"

This bug happens as Y is copied to dat (which also changes X), therefore the model achieves a perfect prediction (100% accuracy).