PLSRegression contains static methods for doing partial least squares regression analysis and prediction of the data.
Constructor | Description |
Full Usage:
PLS2Regression()
|
![]() ![]() ![]() ![]()
|
Instance member | Description |
|
![]() ![]() ![]() ![]()
|
|
![]() ![]() ![]() ![]()
|
|
![]() ![]() ![]() ![]()
|
|
![]() ![]() ![]() ![]()
|
Full Usage:
this.PredictedYAndSpectralResidualsFromPreprocessed
Parameters:
IROMatrix<float>
-
Matrix of unknown spectra (preprocessed the same way as the calibration spectra).
numFactors : int
-
Number of factors used for prediction.
predictedY : IMatrix<float>
-
On return, holds the predicted y values. (They are centered).
spectralResiduals : IMatrix<float>
-
On return, holds the spectral residual values.
Modifiers: abstract |
![]() ![]() ![]() ![]() This predicts concentrations of unknown spectra.
|
Full Usage:
this.Reset
Modifiers: abstract |
![]() ![]() ![]() ![]()
|
Full Usage:
this.SetCalibrationModel
Parameters:
IMultivariateCalibrationModel
Modifiers: abstract |
![]() ![]() ![]() ![]()
|
Static member | Description |
![]() ![]() ![]() ![]()
|
|
Full Usage:
PLS2Regression.CreateFromPreprocessed(matrixX, matrixY, maxFactors)
Parameters:
IROMatrix<float>
-
The spectral matrix (each spectrum is a row in the matrix). They must at least be centered.
matrixY : IROMatrix<float>
-
The matrix of concentrations (each experiment is a row in the matrix). They must at least be centered.
maxFactors : int
-
Maximum number of factors for analysis.
Returns: PLS2Regression
A regression object, which holds all the loads and weights neccessary for further calculations.
|
![]() ![]() ![]() ![]() Creates an analyis from preprocessed spectra and preprocessed concentrations.
|
Full Usage:
PLS2Regression.ExecuteAnalysis(_X, _Y, numFactors, xLoads, yLoads, W, V, PRESS)
Parameters:
IROMatrix<float>
-
The X ("spectrum") matrix, centered and preprocessed.
_Y : IROMatrix<float>
-
The Y ("concentration") matrix (centered).
numFactors : byref<int>
-
Number of factors to calculate.
xLoads : IBottomExtensibleMatrix<float>
-
Returns the matrix of eigenvectors of X. Should be initially empty.
yLoads : IBottomExtensibleMatrix<float>
-
Returns the matrix of eigenvectors of Y. Should be initially empty.
W : IBottomExtensibleMatrix<float>
-
Returns the matrix of weighting values. Should be initially empty.
V : IRightExtensibleMatrix<float>
-
Returns the vector of cross products. Should be initially empty.
PRESS : IExtensibleVector<float>
-
If not null, the PRESS value of each factor is stored (vertically) here.
|
![]() ![]() ![]() ![]() Partial least squares (PLS) decomposition of the matrizes X and Y.
|
Full Usage:
PLS2Regression.GetPredictionScoreMatrix(xLoads, yLoads, W, V, numFactors, predictionScores)
Parameters:
IROMatrix<float>
-
Matrix of spectral loads [factors,spectral bins].
yLoads : IROMatrix<float>
-
Matrix of concentration loads[factors, number of concentrations].
W : IROMatrix<float>
-
Matrix of spectral weightings [factors,spectral bins].
V : IROMatrix<float>
-
Cross product matrix[1,factors].
numFactors : int
-
Number of factors to use to calculate the score matrix.
predictionScores : IMatrix<float>
-
Output: the resulting score matrix[ spectral bins, numberOfConcentrations]
|
![]() ![]() ![]() ![]() Get the prediction score matrix. To get the predictions, you have to multiply the spectras to predict by this prediction score matrix (in case of a single y-variable), this is simply the dot product.
|
Full Usage:
PLS2Regression.Predict(XU, xLoads, yLoads, W, V, numFactors, predictedY, spectralResiduals)
Parameters:
IROMatrix<float>
xLoads : IROMatrix<float>
yLoads : IROMatrix<float>
W : IROMatrix<float>
V : IROMatrix<float>
numFactors : int
predictedY : IMatrix<float>
spectralResiduals : IMatrix<float>
|
![]() ![]() ![]() ![]()
|