Back to PLS HelpPosted on 06/18/16 17:11:51
Number of posts: 2
Dear Colleagues,
Thanks a lot for the toolbox! Could you please suggest an easy strategy for predicting LV scores on a new independent sample?
If you could just wite and example command that would allow projecting new image onto the LV space, it would be particualry great.
Thanks a lot in advance, and sorry if this question is too basic and/or has already been asked elswhere.
Best Wishes,
Alexander
Replies:
Posted on 06/18/16 18:28:50
Number of posts: 394
Dear Colleagues,
Thanks a lot for the toolbox! Could you please suggest an easy strategy for predicting LV scores on a new independent sample?
If you could just wite and example command that would allow projecting new image onto the LV space, it would be particualry great.
Thanks a lot in advance, and sorry if this question is too basic and/or has already been asked elswhere.
Best Wishes,
Alexander
Glad you like the tools. We are hoping to release a significant update in the Fall!
As for your question, in the results structure matrix there is a variable 'u' which contains the singular vector for the brain images (assuming you are doing imaging). The weights in ''u' are what you need to compute scores on a new dataset: newscores= newdata * u;
Login to reply to this topic.
