More Powerful Selective Kernel Tests for Feature Selection

This repository contains resources for selecting statistically significant features using multiscale bootstrap. The test corrests for selection bias. The algorithm is described in our paper,

Lim, J., Yamada, M., Jitkrittum, W., Terada, Y., Matsui, S., Shimodaira, H.
More Powerful Selective Kernel Tests for Feature Selection
AISTATS 2020

How to install?

Requires numpy, matplotlib, SciPy, sklearn. The package can be installed with the following command

pip install git+https://github.com/jenninglim/multiscale-features

Once installed, you should be able to do import mskernel without any error.

Demo

See notebooks.

Reproducing results

See experiments for experiment setup and its corresponding figures can be seen in figures.

See also

• Kernel Multiple Model Comparison: Code Paper.
• Kernel Goodness of Fit (where some of the kernel code is from): Code, Paper