Why PLS is distribution-free?

Moderator: MultiD Support

Why PLS is distribution-free?

Postby joewangtw » Fri Nov 13, 2009 4:14 pm

could anyone tell me, why PLS have distribution-free characteristic ?
I know it's soft model, but what to cause it become distribution-free???
Posts: 1
Joined: Fri Nov 13, 2009 4:05 pm

Re: Why PLS is distribution-free?

Postby Mikael Kubista » Fri Nov 20, 2009 5:15 am

In your data set you have predictors, which are the measure Cq values, and responses, which is a trend variable, such as time, dosage, age etc. When PLS is applied the dimensional spaces, or if you like, the coordinate systems describing the predictors and responses are rotated to give a linear correlation between the two. This process is model independent, although highly skewed distributions may not be possible to bring into a linear relation without severe over fitting.

If you have a good PLS qPCR study that you are ready to share, we would gladly consider to use them as a GenEx example. Let us know if you are interested.

Good luck
Mikael Kubista
Posts: 455
Joined: Tue Jul 01, 2008 12:28 pm

Return to Partial Least Squares (PLS)

Who is online

Users browsing this forum: No registered users and 1 guest


MultiD Analyses

Home of the GenEx analysis software