I use the plsr function in R with cross validation (10-fold). As a result, I get this output: enter image description here

From my limited understanding, I know that the ideal number of components is usually chosen by the minimum adjCV value. Here, the intercept has the lowest value.

I tried finding information for this but didn't find anything. What does this mean?


1 Answer 1


Intercept aka zero component model means: do not fit any PLS components/latent variables but always predict the average training y.

This is a baseline against which the improvement for more complex models can be judged.

In your case, there is no improvement. That observation alone is not sufficient to judge what exactly goes wrong here.

What does the summary tell you about explained x and y variances? What is the y range?


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.