Back to PLS Helpcorrelation values
Posted on 02/25/11 12:58:16
Number of posts: 7
Hi there,
I have just noticed that when I observe brain LV (opposed to bootstrap ratio) the values range beyond +1 to -1 (approx 1.6 to -1.5). Up until now I have presumed these numbers to be r values but evidently they're not due to their range.
If someone could possibly direct me on how to interpret these values, that'd be brilliant!
Thank you, Anna
I have just noticed that when I observe brain LV (opposed to bootstrap ratio) the values range beyond +1 to -1 (approx 1.6 to -1.5). Up until now I have presumed these numbers to be r values but evidently they're not due to their range.
If someone could possibly direct me on how to interpret these values, that'd be brilliant!
Thank you, Anna
Replies:
Untitled Post
Posted on 02/25/11 16:31:49
Number of posts: 394
The values are weights within the singular vector (think eigenvector) reweighted by the the singular value (think eigenvalue). As such, they are only relevant as a means to evaluate the weights between regions within the same latent variable. You can make a link between these numbers of unstandardized regression weights, so there is no theoretical limit since the values depend on the scale of measure.
Login to reply to this topic.
