Bug in KernelPCA?

hieukienghieukieng Member Posts: 1 Contributor I
edited November 2018 in Help
Sr if this is the inappropriate place to post this kind of information.
As I tried to use KernelPCA with my classification problem, I could not apply trained model to test instances. The message "The example set contains attributes in different order than on creation time...." appeared although the attributes set are exactly the same. It turned out from the source code that when I changed this line of code in method checkNames() of class KernelPCAModel
...
if (attribute.getName() != attributeNames.get(i++)) {
...
to
...
if (!attributeNames.get(i++).equals(attribute.getName())) {
...
so I could applied train model for test instances. It really helped me.
I am not sure whether this is a bug or not.
Hope that help!

Answers

  • landland RapidMiner Certified Analyst, RapidMiner Certified Expert, Member Posts: 2,531 Unicorn
    Hi,
    this is definitively a bug, since Strings are not the same even if they equal. I corrected this.

    Thank you,
    Β  Sebastian
Sign In or Register to comment.