mcorr

%MCORR Multi-plot of all correlations between columns of a matrix
% MCORR (X) plots correlations between all possible combinations of the
% columns of array X, in a single figure. If the first argument is the
% name of a file in the current directory, mcorr reads
% it (including variable names in the first row) as X. Otherwise it is
% assumed that the first argument is an array, in which case consecutive
% numbers will be used as variable names. If there is a second (numeric)
% argument, mcorr will plot only the columns indicated in the second argument.
% The frequency distribution of each variable is plotted in the main
% diagonal with HIST using n/20 bins.
% It is unpractical to try to plot more than, say, 8 variables, since each
% individual plot becomes too small.
%
% OUTPUT= MCORR (X,'sig',ALPHA) calculates the correlation coefficient between
% each pair of columns and, if the correlation is sgnificant at the ALPHA
% level, points are plotted in red and the column numbers, corr. coefficient
% and p-value are returned in the OUTPUT array. Requires Statistical Toolbox
%
% OUTPUT= MCORR (X,...,'corr',CORTYPE) specifies the type of correlation,
% allowed values are 'Pearson' (default), 'Spearman' and 'Kendall'
%
% EXAMPLES:
% mcorr ('myfile')
% mcorr ('myfile',[1:5])
% mcorr (X,[3:6])
% output= mcorr (X,'sig',0.05)
%
% Last modified: Jun. 2012