<div dir="ltr"><pre>Hello everyone,<br><br>I have a problem about components selection for SIFT. Since I am working
with ICA components, so I use ADJUST or MARA to reject the noisy
components. However, I need to select few components in each region and
remove the unnecessary ones. , do you know any other good ways to do this
work?
According to the SIFT handbook, it said to combined two approaches which
are MARA or ADJUST to remove bad components and estimating the partial
coherence between the all the variables using non-parametric methods (e.g.
FFT or wavelet). I am not quite sure how to manage to do it. Would you mind
share your idea?<br><br></pre><pre>Cheers,<br></pre><pre>MO XIAO<br></pre></div>