pls()

pls(A,Y,n)

Fit matrix A and matrix/vector Y using the PLS technique; matrix A contains more than one row and both A and Y should have same number of rows; n is the number of principal components. The type of result is determined by Y, that is to say, the function returns a matrix when Y is a matix and a vector when Y is a vector. The return result can be used as parameter F in pls(A’,F) to perform a fitting

pls(A’,F)

Get fitting object of matrices having same number of columns, that is, perform prediction on A’ using coefficient information in a fitting result

A

A matrix

Y

A matrix/vector

n

Number of principal components; the range of its values is 1<=n<=min(row count -1,column count). When the parameter is absent or out of range, its value will be automatically configured as the maximum in the above range

A’

A matrix/vector

F

A matrix/vector, the result returned by pls(A,Y,n)

A

1

[[4,6,2,4],[2,3,1,2],[1,1,1,-1]]

2

[1,0,-4]

Here parameter Y is a vector.

3

=pls(A1,A2,2)

Fit matrix A1 and vector A2 using the PLS technique and return a vector; the number of principle components is 2.

4

[[3,5,-2,8],[2,4,3,3]]

5

=pls(A4,A3)

Perform prediction on A4 using fitting result returned by A3.

6

[[1,-2,1],[0,2,-8],[-4,5,9]]

Here parameter Y is a matrix.

7

=pls(A1,A6,2)

Fit together matrix A1 and matrix A6 using PLS technique, with the principle component is 2, and return the fitting coefficient matrix.

8

=pls(A4,A7)

Perform prediction on A4 using fitting result returned by A7 and return prediction result.