PLS2: multi-output PLS and its outputs - % explained for each output and alternatives
11 views (last 30 days)
Show older comments
Dear community,
I want to build multi-output PLS model (PLS2) via plsregress (built-in function). From what I see in the documentation, MATLAB implements PLS1 (1 PLS model per output) or PLS2 (a single PLS model for all outputs) simply depending on the size of the response variable vector/matrix supplied. However, I am also interested in the function output PCTVAR, but it only shows me the % of the variance explained in the data and in the overall response matrix. Two questions arise:
1 - Is there a way to know the % explained per output?
2 - If not PLS2 via plsregress, would you recommend another alternative (function/technique) for a method that simultaneously does data reduction and regression for correlated outputs? (needs to be MIMO, application: spectral processing)
From the documentation, verbatim:
Y : "Response variables, specified as a numeric matrix. Y is an n-by-m matrix, where n is the number of observations and m is the number of response variables."
PCTVAR: "Percentage of variance explained by the model, returned as a numeric matrix. PCTVAR is a 2-by-ncomp matrix, where ncomp is the number of PLS components. The first row of PCTVAR contains the percentage of variance explained in X by each PLS component, and the second row contains the percentage of variance explained in Y."
0 Comments
Answers (0)
See Also
Categories
Find more on Dimensionality Reduction and Feature Extraction in Help Center and File Exchange
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!