-------------------------------------------
chi-square Frank Test
-------------------------------------------
Author: Arno Onken <aonken@cs.tu-berlin.de>

This package provides a test for linear dependence structure between two count vectors x and y. For conceptual details see Onken, Gruenewaelder, and Obermayer 2009, Advances in Neural Information Processing Systems 22.

The main function is called "chi2franktest".
 Arguments:
  x         - Vector of integer random values (NumPy array for Python)
  y         - Vector of integer random values of the same size as x (NumPy array for Python)
  alpha     - Significance level (default alpha = 0.05)
  mef       - Minimum expected frequency (default mef = 1)
 Returns:
  h         - True indicates rejection of the linear dependence hypothesis at the specified significance level; False otherwise
  thresh    - chi-square threshold value
  test_stat - Test statistic

The software was tested in:
 MATLAB 7.6.0
 GNU Octave 3.0.5
 Python 2.5.4 with NumPy and SciPy modules

