Create a Correlation matrix from the selected TableOfReal and the GaussianMixture.
We start by calculating the ClassificationTable from the data and the GaussianMixture.
Nex we calculate correlations between the columns of the ClassificationTable: cell [i,j] of the correlation matrix will then contain the value:
|p[i] · p[j] / (||p[i]||· ||p[j]||),|
where p[i] is the data in the i-th column of the classification table and ||p[i]|| is its Euclidean norm. The index i runs from 1 to the number of components in the mixture.
Because all the elements in a column are positive numbers, i.e. probabilities, all correlations will be positive numbers too.
© djmw, November 11, 2010