Why pca doesn't work on matlabR2013b?

1 view (last 30 days)
Michelle
Michelle on 11 Mar 2014
Commented: Anne on 11 Jul 2014
I'm trying to run pca on a dataset that I have and I'm finding an error, then I tried:
TT=pca(randn(20,200));
It gives me the same following error message:
Undefined variable "classreg" or class "classreg.learning.internal.wnanmean".
Error in pca (line 297) mu = classreg.learning.internal.wnanmean(x, vWeights);
  4 Comments
Anne
Anne on 11 Jul 2014
If anyone else has this problem in the future, I just got it resolved through tech support.
It seems like MATLAB might be looking in the wrong directories for the definition of "classreg". This error occurs during installation sometimes. Please execute the following at the command line:
>> restoredefaulthpath; >> rehash toolboxcache; >> savepath;
Then, try executing your code.

Sign in to comment.

Answers (2)

Youssef  Khmou
Youssef Khmou on 11 Mar 2014
Edited: Youssef Khmou on 11 Mar 2014
hi, i am not aware about the error, but technically this alternative multiscale pca needs some details , let us take the example you tried before, 20 signals each of length 200.
x=randn(200,20);
% additional parameters.
level = 5;
wname = 'sym4'; %
npc = 'kais';
[x_sim, qual, npc] = wmspca(x,level,wname,npc);
subplot(1,2,1); plot(x(:,1)); title(' original first signal')
subplot(1,2,2); plot(x_sim(:,1),'r'); title(' transformed first signal')
try to manipulate the parameters using Documentation
But that result damaged the correlation property of the random signals :
surf(cov(x))
figure; surf(cov(x_sim))
  2 Comments
Youssef  Khmou
Youssef Khmou on 11 Mar 2014
try this function if you cant fix the error .
<http://crcv.ucf.edu/source/dimension%20reduction/pca.m>

Sign in to comment.


Ilya
Ilya on 11 Mar 2014
I suggest that you get in touch with the tech support. (You can find their phone number and email at mathworks.com.) There may be an issue in your MATLAB installation.

Tags

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!