Partial least squares confusion

Hi All,

I am having some problems getting my head around a PLS method from a
paper by Bastien et al. I can get as far as creating the term t1, yet I
have as yet been unsuccessful in creating terms t2, t3 etc... and I am
looking for some assistance.

I have put links to the paper and to images of what I believe are the
relative sections at the end of this post so you can view the method I
am working with (the example begins on pg 25 of the journal)

The example given based on the following data set

% Y X1 X2 X3 X4 X5 X6 X7
data = [98.70 0.00 0.23 0.00 0.00 0.00 0.74 0.03;...
97.80 0.00 0.10 0.00 0.00 0.12 0.74 0.04;...
96.60 0.00 0.00 0.00 0.10 0.12 0.74 0.04;...
92.00 0.00 0.49 0.00 0.00 0.12 0.37 0.02;...
86.60 0.00 0.00 0.00 0.62 0.12 0.18 0.08;...
91.20 0.00 0.62 0.00 0.00 0.00 0.37 0.01;...
81.90 0.17 0.27 0.10 0.38 0.00 0.00 0.08;...
83.10 0.17 0.19 0.10 0.38 0.02 0.06 0.08;...
82.40 0.17 0.21 0.10 0.38 0.00 0.06 0.08;...
83.20 0.17 0.15 0.10 0.38 0.02 0.10 0.08;...
81.40 0.21 0.36 0.12 0.25 0.00 0.00 0.06;...
88.10 0.00 0.00 0.00 0.55 0.00 0.37 0.08];

I find significant terms are X1, X3, X4, X6 & X7 which gives a t1 term

t1 = 0.4526(x1*) - 0.4530(x3*) - 0.3820(x4*) + 0.5325(x6*) -
0.4006(x7*)

The next instruction is where I get stuck

"Computation of t2: At first, the variables Xj's with a significant
contribution to the construction of t2 are looked for. Therefore, 7
multiple regressions of y* on t1 and each Xj are run".

The significance levels (p-values) are reported as

X1 0.6225
X2 0.0101
X3 0.6016
X4 0.9055
X5 0.7221
X6 <0.0001
X7 0.0532

I can not get the same p-values which means I am doing something wrong
in completing the multiple regressions of y* on t1 and each Xj.

I would be extremely grateful if someone can help me with this. I have
worked at it for a while now with no success and I think it is
relatively straight forward for a stats expert.

Kind regards

Dan

LINKS:
The paper by Bastien et al
http://www.danpearce.co.uk/download/pls_glr.pdf

Relevant sections
http://i21.photobucket.com/albums/b266/dp_pearce/pic1.gif
http://i21.photobucket.com/albums/b266/dp_pearce/pic2.gif

.

Relevant Pages

  • Partial least squares confusion
    ... The significance levels are reported as ... I can not get the same p-values which means I am doing something wrong ... in completing the multiple regressions of y* on t1 and each Xj. ... relatively straight forward for a stats expert. ...
    (sci.stat.math)
  • Re: The absence of evidence is not evidence of absence
    ... The reader is reminded that, among other incredibly stupid things, Afonso does not know the difference between significance levels and p-values: ... These are p-values, not significance levels. ...
    (sci.stat.math)