error using tranpose (recieving permute error )

CODE:

%% Testing Phase
load('testFeatures');
featvales=b2{1};
label=b2{2};

num_dims = max(size(featvales));

X=featvales';
Y=label';

PROBLEM:

I have to calculate the average for 2 different dimension

1st for 10 dimensional function

2nd for 30 dimensional function

ERROR :(the following error is recieved while executing the code)

Error using '

Transpose on ND array is not defined. Use PERMUTE instead.

Error in FeatureSelection (line 28)

X=featvales';

[Walter Roberson]

num_dims = max(size(featvales));

That is incorrect.

num_dims = ndims(featvales);

[Meetul Choudhary]

still getting the same error

Answer 1

[Fangjun Jiang]

The error message is clear. How do you transpose a three (or more) dimension array?

a=rand(2,3);
a';
a=rand(2,3,4);
a';

Error using '
TRANSPOSE does not support N-D arrays. Use PAGETRANSPOSE/PAGECTRANSPOSE to transpose pages or PERMUTE to reorder dimensions of N-D arrays.

[Meetul Choudhary]

Error using '

Transpose on ND array is not defined. Use PERMUTE instead.

Error in FeatureSelection (line 29)

X';

I'm still getting this error. Is there a way we can connect and discuss this code?

[Fangjun Jiang]

What is the return value of size(X)?

Your array "X" must be a multi-dimenstion array. In that case, you can't do transpose(X).

[Meetul Choudhary]

[Fangjun Jiang]

add this line before doing the transpose. You will see it removes the two dimensions of length 1.

featvales=squeeze(featvales);
size(featvales)

[Meetul Choudhary]

[Fangjun Jiang]

Then you should be able to do featvales'

[Meetul Choudhary]

can you connect to me personally @[e-mail]

Answer 2

[James Tursa]

If you are trying to transpose the initial 2D pages of your variable, you can do this:

pagetranspose(featvales)

Or this:

d = 1:ndims(featvales)
permute(featvales,[2,1,d(3:end)])

But it isn't clear if this will fix your underlying problem.

[Meetul Choudhary]

what exactly does this output gives as output?

it is showing a value for (:, :, 1, 704) to (:, :, 1000, 704)

[Meetul Choudhary]

I am working on feature selection where im facing this issue in a part of code