KOLMOGOROV_SMIRNOV computes the two-sample Kolmogorov-Smirnov test for 
   each pair columns. If data size does not match, the data can be 
   filled up with not-a-number (NaN).

 Usage: 
    [D,ks,pval,df] = kolmogorov_smirnov(x,y);
    [...] = kolmogorov_smirnov(X);
    [...] = kolmogorov_smirnov(..,'Tail',tail);

 Input: 
    x,y input vectors for comparison
    X   data matrix, each column represents a sample distribution
        in case, the number of samples do not match, the matrix can
        can be filled up with not-a-number (NaN) values. 
    tail: 'unequal' (default), 'larger', 'smaller'

 Output:
    D   maximum absolute difference between sample data
        D(k,l) is the m.a.d. from X(:,k) and X(:,l) 

    df  is the degree-of freedom 
        df(k,l) = n(k)*n(l)/(n(k)+n(l)) with n samples of corresponding 
        column X. 
    pval  p-value, it's also a matrix where 
        pval(k,l) is the p-value from column k and l. 

  NOTE: For M data sets (M-columns of X), there are M*(M-1)/2 tests, 
    you might want to consider some correction for multiple 
    comparision. 
 
 see also:
    kstest2, statistic/kolmogorov_smirnov_test_2

Package: nan