3

There's a Wiki page for PLS "Path Modeling" and another one for PLS "Regression"... they seem to be saying the same thing, except the first link says PLS is great, and the second says that PLS is controversial and some journals no longer accept papers based on PLS:

1) https://en.wikipedia.org/wiki/Partial_least_squares_path_modeling

2) https://en.wikipedia.org/wiki/Partial_least_squares_regression

Are PLS "Path Modeling" and PLS "Regression" pretty much the same thing?

thanks_in_advance
  • 735
  • 1
  • 7
  • 14
  • 2
    I don't exactly know what is meant by path modelling. I understand, however, from what I read that the PLS algorithm is the same for both cases whereas their usage areas differ. PLS is a fantastic tool for the people dealing with spectra, and there is a reason why it is so popular: it works for many many cases. IMHO, there is almost nothing controversial about PLS in the field of chemometrics, however the very same method may contradict with the nature of other type of data or procedure. – gunakkoc Feb 13 '17 at 07:36
  • 1
    +1, I have been wondering about this issue before. Would really like to see this question answered. Unfortunately I am not familiar with meaning #1 either, so cannot really answer myself. @theGD: "path modeling" refers to https://en.wikipedia.org/wiki/Structural_equation_modeling. – amoeba Feb 13 '17 at 11:43
  • @theGD path modeling is like structural equation modeling with no latent variables. Causal structures can be tested in complex graphical models with several possible mediating or moderating effects handled slightly different from the case in linear regression. – AdamO Feb 13 '18 at 21:52

0 Answers0