/HSIC

Python code of Hilbert-Schmidt Independence Criterion

Primary LanguageMATLABMIT LicenseMIT

Hilbert-Schmidt Independence Criterion (HSIC)

Python version of the original MATLAB code of Hilbert-Schmidt Independence Criterion (HSIC).

Prerequisites

  • numpy
  • scipy

We tested the code using Anaconda 4.3.0 64-bit for python 2.7 on windows.

Apply on your data

Usage

Import HSIC using

from HSIC import hsic_gam

Apply HSIC on your data

testStat, thresh = hsic_gam(x, y, alph = 0.05)

Description

Input of function hsic_gam()

Argument Description
x Data of the first variable. (n, dim_x) numpy array.
y Data of the second variable. (n, dim_y) numpy array.
alph level of the test

Output of function hsic_gam()

Argument Description
testStat test statistic
thresh test threshold for level alpha test

Independence test result

  • If testStat < thresh, x and y are independent.
  • If testStat > thresh, x and y are not independent.

Authors

  • Shoubo Hu - shoubo [dot] sub [at] gmail [dot] com

See also the list of contributors who participated in this project.

License

This project is licensed under the MIT License - see the LICENSE file for details.