PLS and PCA

February 7, 2002, 05:15

Hi,

I am hunting for good resource on Partial Least Square (PLS) Algorithm and principal Component Analysis (PCA).

There are many sites which provide information about it. I could not locate a single one which illustrates the algorithm depicting its pseudo code. Typically, I am hunting for pseudo code of these algorithms.

Any resource or pointer in this regard is highly appreciable.

Thanks in advance

Pargaonkar Vijay

February 7, 2002, 07:34

Can't help you with the PLS but for the PCA stuff do a reference search on Proper Orthogonal Decompostion(POD) and CFD...in reality the algorithm for POD is just 1)Form a matrix which contains snapshots in time 2) Form the correlation matrix 3)Find the eigenvalues of the correlation matrix 4)Pick n (where n is much smaller than the original # of DOF) of these eigenvectors and assume the dof are a linear combination of these eigenvectors

February 8, 2002, 05:19

The University of Newcastle have a forum called INPACT - you may wish to search for this.

Otherwise, try code vendors web sites for links - these include UMETRICS, SIMCA-P.

February 7, 2002, 05:15	PLS and PCA	#1
Pargaonkar Vijay Guest Posts: n/a	Hi, I am hunting for good resource on Partial Least Square (PLS) Algorithm and principal Component Analysis (PCA). There are many sites which provide information about it. I could not locate a single one which illustrates the algorithm depicting its pseudo code. Typically, I am hunting for pseudo code of these algorithms. Any resource or pointer in this regard is highly appreciable. Thanks in advance Pargaonkar Vijay

February 7, 2002, 07:34	Re: PLS and PCA	#2
Peter Attar Guest Posts: n/a	Can't help you with the PLS but for the PCA stuff do a reference search on Proper Orthogonal Decompostion(POD) and CFD...in reality the algorithm for POD is just 1)Form a matrix which contains snapshots in time 2) Form the correlation matrix 3)Find the eigenvalues of the correlation matrix 4)Pick n (where n is much smaller than the original # of DOF) of these eigenvectors and assume the dof are a linear combination of these eigenvectors

February 8, 2002, 05:19	Re: PLS and PCA	#3
gnu Guest Posts: n/a	The University of Newcastle have a forum called INPACT - you may wish to search for this. Otherwise, try code vendors web sites for links - these include UMETRICS, SIMCA-P.