I have implemented PCA on training data and the result is ready. Now I want to use the same PCA on test. Is there any way to perform the same PCA on test as training in weka?
$\begingroup$ $\endgroup$
2 
- $\begingroup$ What do you mean "test" here? Cases that did not participate in the PCA and for which you need to compute PC scores? $\endgroup$ttnphns– ttnphns2014-04-13 07:51:54 +00:00Commented Apr 13, 2014 at 7:51
- $\begingroup$ the training data are which has class label and test do not. I have performed pca on training now, with the same parameter I wanna do it for test. for example if the first attribute in training is "0.2*h1+0.7*h2" I wanna have this attribute for test. $\endgroup$Paniz– Paniz2014-04-14 13:55:50 +00:00Commented Apr 14, 2014 at 13:55
Add a comment |
1 Answer
$\begingroup$ $\endgroup$
This is easily done by using batch filtering, that allows applying a filter to both a training and a test set in WEKA. Your only problem is that you will have to apply PCA to the training set again -- although you should get the same results if you make use of the same options.
