# how do i calculate the Correlation coefficients

47 views (last 30 days)
neamah al-naffakh on 12 Aug 2016
Commented: the cyclist on 13 Aug 2016
I have dataset that contains the acceleration data of three axes (x, y, and z) , this an example of my data
X Y Z
%1.0195313 0.16088867 -0.26391602
%1.0976563 0.17456055 -0.33447266
%1.2556152 0.24926758 -0.3774414
%1.2314453 0.517334 -0.25732422
%1.0212402 0.5761719 -0.09277344
%1.0727539 0.51000977 0.007324219
%1.1694336 0.32885742 -0.017822266
%1.1247559 0.22924805 -0.10595703
%1.0339355 0.27905273 -0.13623047
%0.8273926 0.24560547 -0.080566406
%0.75097656 0.25390625 -0.018310547
%0.67626953 0.3046875 0.051757813
%0.7282715 0.33764648 0.08911133
%0.8227539 0.34545898 0.08325195
%0.87939453 0.3413086 0.08300781
%1.0527344 0.34326172 0.106933594
%1.1125488 0.32128906 0.10058594
%1.1477051 0.26416016 0.06347656
%1.2807617 0.2680664 0.015136719
%1.2800293 0.2668457 -0.006347656
I have calculated the Correlation coefficients of two axes (x and y) using the corrcoef function as illustrated bellow
R = corrcoef(X_Segments{1},Y_Segments{1})
X_Segments contains data of the first column (X Data) and Y_Segments contains data of the second column (Y Data). However, the output was strange
R =
1.0000 0.3424
0.3424 1.0000
I want to know which value represents the Correlation coefficients between X and Y and is there any solution to get the Correlation coefficients between two axes directly as one value (single output) not as a matrix ?

neamah al-naffakh on 13 Aug 2016
Edited: neamah al-naffakh on 13 Aug 2016
Hi guys, thanks for your answer howeover, the correct function is corr
CorrcoeffXY{nn,1} = corr(X_Segments{nn},Y_Segments{nn});
and the output will be the Correlation coefficients between X and Y
R = 0.3424
the cyclist on 13 Aug 2016
Oh, of course. I only tried
corr(rand(5,2))
(which returns a matrix)
and not
corr(rand(5,1),rand(5,1))
(which returns a scalar).
Glad you found something that worked for you.

the cyclist on 13 Aug 2016
The upper-right and lower-left elements are equal, and each one is the correlation you want. The diagonal elements are the trivial correlation of X with itself and Y with itself. Those will always be equal to 1, by construction.
I don't think there is a simple way to get just the single number without first getting the matrix.
Image Analyst on 13 Aug 2016
So, (just to be super explicit)
correlationBetweenXandY = R(1,2);
Is that what you want neamah?