Back to PLS Helpautocorrelations in fmri data
Posted on 02/02/07 12:15:28
Number of posts: 100
uthor: Jennifer Labus (149.142.243.---)
Date: 11-29-06 18:09
How does PLS account for autocorrelations in the data? Does it and if not why is this a good strategy?
Date: 11-29-06 18:09
How does PLS account for autocorrelations in the data? Does it and if not why is this a good strategy?
Replies:
Re: autocorrelations in fmri data
Posted on 02/02/07 12:15:49
Number of posts: 100
Author: Randy (---.rotman-baycrest.on.ca)
Date: 11-30-06 17:12
> How does PLS account for autocorrelations in the data? Does
> it and if not why is this a good strategy?
It does not account for it in any formal way. For most analyses the autocorrelation impacts the derivation of the appropriate error term, which does not exist in the inferential assessment of the statistical patterns in PLS.
In another sense, the fact that there is both a spatial and temporal dependency in the data is exactly why you want to use PLS to take advantage of the inherent "redundancy" in the data to better capture the network-level response.
Date: 11-30-06 17:12
> How does PLS account for autocorrelations in the data? Does
> it and if not why is this a good strategy?
It does not account for it in any formal way. For most analyses the autocorrelation impacts the derivation of the appropriate error term, which does not exist in the inferential assessment of the statistical patterns in PLS.
In another sense, the fact that there is both a spatial and temporal dependency in the data is exactly why you want to use PLS to take advantage of the inherent "redundancy" in the data to better capture the network-level response.
Login to reply to this topic.
