Home > brede > brede_mat_partcorrcoef_test.m

brede_mat_partcorrcoef_test

PURPOSE ^

brede_mat_partcorrcoef_test - Correlation coefficient

SYNOPSIS ^

function P = brede_mat_partcorrcoef_test(X, Y, Z, varargin)

DESCRIPTION ^

 brede_mat_partcorrcoef_test  - Correlation coefficient

       P = brede_mat_partcorrcoef_test(X, Y, Z, varargin)

       Input:    X    'Mat' structure
                 Y    'Mat' structure
                 Z    'Mat' structure with nuisances

       Property: Info          [ {0} | integer | Waitbar ] Amount of
                               running/debug information
                 MaxIterations [ {10000} | integer ] maximum number
                               of Monte Carlo permutations for the
                               permutation test
                 Seed          [ {0} | a value ] Seed for the random
                               generation in the Monte Carlo permutation
                               test 
                 Type          [ {TwoTailed} | LowerTail |
                               UpperTail | MaxPermuteTwoTailed ]
                               Type of test 

       Output:   P    'Mat' struture with P-values

       Test with univariate partial correlation coefficient between X
       and Y with Z as the nuisance/covariates. 

       This test can also be performed from a graphical user
       interface via the brede_ui_mat_partcorrcoef function. 

       'TwoTailed' will perform a test with Gaussian assumptions
       using a t-distribution. The resulting P-value will not be
       corrected for multiple comparisons.

       Type='UpperTail' will perform a one-tailed test against the
       t-distribution, i.e., measure the upper tail area of this
       distribution.

       Type='MaxPermuteTwoTailed' will make multiple correction with
       the permutations and the maximum statistics. 'MaxIterations'
       specifies the maximum number of Monte Carlo permutations to
       perform. 

       Example: 
         % Data without correlation
         P = brede_mat_partcorrcoef_test(randn(20,3), randn(20,4000), ...
               randn(20,2), 'type', 'UpperTail');
         figure, hist(P.matrix(:), 100) 

         % Positive correlation data
         X = randn(20,100);
         P = brede_mat_partcorrcoef_test(X, randn(20,100)+X, ...
               randn(20,2), 'type', 'UpperTail');
         figure, hist(diag(P.matrix), 10)
 
       See also BREDE, BREDE_MAT, BREDE_MAT_CCA, BREDE_MAT_CORRCOEF,
                BREDE_MAT_COV, BREDE_MAT_GLM, BREDE_MAT_MEAN,
                BREDE_MAT_ORTHOSPACE, BREDE_MAT_PARTCORRCOEF,
                BREDE_MAT_PARTCORRCOEF_TVALUE, BREDE_CDF_R,
                BREDE_CDF_T, BREDE_UI_MAT_PARTCORRCOEF.

 $Id: brede_mat_partcorrcoef_test.m,v 1.16 2009/09/25 12:40:36 fn Exp $

CROSS-REFERENCE INFORMATION ^

This function calls: This function is called by:
Generated on Fri 27-Nov-2009 18:11:22 by m2html © 2005