Dear,

I am solving the problem of Partial least Square Regression problem for the Human Whole blood for 7 varients( Glucose, Urea, Cholestrol etc). I have spectra for 288 samples having different concentration of above 7 varients. The data points we are generating for each sample are dependent on the wavelength having range from 2.0 - 2.4 nm with the interval of 0.005 nm. Hence for each sample we have total 101 datapoints.

My lookout is to compute unknown sample varient concentrations for the spectra of the sample.

I am using Partial Least Square Regression to build the model ( I have been user of the following portal for computation of Partial Least

Square Algorithm for my studies.

http://www.statsoft.com/textbook/stpls.html#basic
I am solving the multivarient system and am using SIMPLE algorithm as

mentioned at

http://www.statsoft.com/textbook/stpls.html#SIMPLS
)

equation Y=XB+ E there

mentioned in the SIMPLE algorithm.

Further kindly infome me how to compute the "X" unknown matrix for the known "Y"

.

I tried solving the equation in matlab like follows.

I computed the matrix 'B' load vectors from the known spectra template

B=W*Q'

This is my unkown sample spectra,

Yun=[2.796798e+004 2.803789e+004 2.810781e+004 2.817773e+004 2.824765e+004 2.831757e+004 2.838749e+004 2.845741e+004 2.852733e+004 2.859725e+004 3.327967e+005 3.336084e+005 3.344201e+005 3.352318e+005 3.360435e+005 3.368552e+005 3.376669e+005 3.384786e+005 3.392903e+005 3.401020e+005 3.409137e+005 3.417254e+005 3.425371e+005 1.184061e+006 1.186861e+006 3.449722e+005 3.457839e+005 3.465956e+005 3.474073e+005 3.482190e+005 3.490307e+005 3.498424e+005 3.506541e+005 3.514658e+005 3.522775e+005 3.530892e+005 3.539009e+005 3.547126e+005 3.555243e+005 3.563360e+005 3.571477e+005 3.579594e+005 3.587711e+005 3.595828e+005 3.603945e+005 3.612062e+005 3.620179e+005 3.628296e+005 3.636413e+005 3.644530e+005 3.652647e+005 3.660764e+005 7.402979e+005 7.419357e+005 3.721096e+005 3.729293e+005 3.737489e+005 3.745685e+005 3.753881e+005 7.207802e+005 7.223505e+005 7.239208e+005 7.254911e+005 7.270615e+005 2.031518e+005 2.035896e+005 2.040274e+005 2.044653e+005 2.049031e+005 2.053409e+005 2.057787e+005 2.062166e+005 2.066544e+005 2.070922e+005 2.075301e+005 2.079679e+005 2.084057e+005 2.088435e+005 2.092814e+005 2.097192e+005 1.359702e+004 1.362535e+004 1.365367e+004 1.368200e+004 1.371033e+004 1.373866e+004 1.376698e+004 1.379531e+004 1.382364e+004 1.385196e+004 1.388029e+004 1.390862e+004 1.393695e+004 1.396527e+004 1.399360e+004 1.402193e+004 1.405025e+004 1.407858e+004 1.410691e+004 1.413524e+004 1.416356e+004 ;

];

Xun=Yun/B

I tried solving the same using the matlab but have not yield satisfactory results.Kindly help me.

Regards

Rajendra Gad,

Goa India