Fisher's exact test with R×C contingency table

This function is based on Monte Carlo simulation, it support contingency tables larger than 2 * 2.

The function 'fishertest' provided by MATLAB only supports contingency table of size 2 × 2, and although there have been many other scholarly published functions on File_Exchange to compute larger matrices, the results are somewhat problematic.

This function is based on Monte Carlo simulation, it support contingency tables larger than 2 * 2.

% Input variable: x

% x is the contingency table. You can use function 'crosstab' to generate the contingency table.

% num : the number of Monte Carlo simulations to be performed, default is 100000.

% Output variable: p

% p：p-values

% Source of subfunction 'rcont2': https://people.math.sc.edu/Burkardt/m_src/asa144/asa144.html

% for example

% x = [ 11 11

% 0 6

% 11 17];

% p = fisher_test_R_C(x)

% p = 0.0977;

% Because of the randomness in the algorithm, the result may be different each time.

% You can increase the value of 'num' to make the result more robust.

Reference:

1. James Boyett,Algorithm AS 144: Random R x C Tables with Given Row and Column Totals,Applied Statistics,Volume 28, Number 3, 1979, pages 329-332.

https://people.math.sc.edu/Burkardt/m_src/asa144/asa144.html