PLSR function

Partial Least Squares Regression

Partial Least Squares Regression for numerical variables.

PLSR(Y, X, S = 2, InitTransform = 5, grouping = NULL, 
centerY = TRUE, scaleY = TRUE, tolerance = 5e-06, 
maxiter = 100, show = FALSE, Validation = NULL, nB = 500)

Arguments

• Y: Matrix of Dependent Variables
• X: Matrix of Independent Variables
• S: Dimension of the solution
• InitTransform: Initial transformation of the independent variables.
• grouping: Fator when the init transformation is the standardization with the within groups deviation.
• centerY: Should the dependent variables be centered?
• scaleY: Should the dependent variables be standadized?
• tolerance: Tolerance for the algorithm
• maxiter: Maximum number of iterations
• show: Show the progress of the algorithm?
• Validation: Validation (None, Cross, Bootstrap)
• nB: number of samples for the bottstrap validation

Details

Partial Least Squares Regression for numerical variables.

Returns

An object of class plsr with fiends

• Method: PLSR

• X: The X matrix

• Y: The Y matrix

• centerY: Is the Y matrix centered

• scaleY: Is the Y matrix scaled

• Initial_Transformation: Initial transformation of the Y matrix

• ScaledX: Transformed X matrix

• ScaledY: Transformed Y matrix

• Intercept: Intercept of the model

• XScores: Scores for the individals from the X matrix

• XWeights: Weigths for the X set

• XLoadings: Loadings for the X set

• YScores: Scores for the individals from the Y matrix

• YWeights: Weigths for the Y set

• YLoadings: Loadings for the Y set

• RegParameters: Final Regression Parameters

• ExpectedY: Expected values of Y

• R2: R-squared

• XStructure: Relation of the X variables with its structure

• YStructure: Relation of the Y variables with its structure

• YXStructure: Relation of the Y variables with the X components

References

H. Abdi, Partial least squares regression and projection on latent structure regression (PLS regression), WIREs Comput. Stat. 2 (2010), pp. 97-106.

Author(s)

Jose Luis Vicente Villardon

See Also

Biplot.PLSR

Examples

X=as.matrix(wine[,4:21])
y=as.numeric(wine[,2])-1
mifit=PLSR(y,X, Validation="None")