17.7.2.1 The Partial Least Squares Dialog Box

Contents

Input

Select data for Partial Least Squares.

Independent Variables Select data to specify the independent variables
Dependent Variables Select data to specify the dependent variables.
Observation Labels Select data to specify label for observation, the labels will be show in the Score Plots and X-Y Scores Plots in result. Note that the label column will be set as categorical if Text column.
Predict Response Determine whether to predict response from the established parital least square model. If checked, Independent Variables for Prediction control is shown.
Independent Variables for Prediction Select data to specify independent variables for prediction

Settings

Specify the settings in Partial Least Squares.

Method Select the method to compute the extracted factors:
  • SVD
  • Wold's Iteration

Note: To learn more details for these two methods, see the algorithm page

Scale Variables Select the method to divide independent variables by the standard deviation
Maximum Number of Factors Specify the maximum number of factors to be in the regression model
Cross Validation Specify whether to establish the partial least square regression model using the Cross Validation method.

Quantities

Descriptive Statistics Specify whether to perform basic statistics(N in analysis, N missing, Mean, Standard Deviation) for each independent and dependent variables. The results will be put in PLSResultsn sheet
Coefficients Specify whether to output the coefficents of the fitted model. The coefficients will be put in PLSResultsn sheet
Standardized Coefficients The option is only available when Scale Variables check box is selected. Specify whether to output the coefficients for the scaled variables. The results will be put in PLSResultsn sheet
Percent Variance Explained for Each X
Percent Variance Explained for Each Y
 Specify whether to output the percent variance for each X or Y variables. The results will be put in PLSResultsn sheet
VIP Specify whether to ouput the variable influence on projections statistics. The results will be put in PLSResultsn sheet
Loadings Specify whether to ouput the loadings for each X and Y variables. The results will be put in PLSLoadingsn sheet
Scores Specify whether to ouput the scores for each X and Y variables. The results will be put in PLSScoresn sheet
Predicted Response for Training Data Speciify whether to output the fitted values. The results will be put in PLSResultsn sheet
Residuals Speciify whether to output the residuals. The results will be put in PLSScoresn sheet
Distance Speciify whether to output the distance to X/Y model for each observation. The results will be put in PLSResultsn sheet
T Square for X Scores Speciify whether to output the t square statistics for each observation. The results will be put in PLSResultsn sheet
X Weights Speciify whether to output the weights for each X variable. The results will be put in PLSLoadingsn sheet

Plots

Specify whether to show plots in the Partial Least Squares. The plots will be added to the PLSn sheet

Coefficients Plot Specify whether to show the Scree Plot for eigenvalues.
Variance Importance Plot Specify whether to show the Variance Importance Plot
Component Plot Specify whether to show plots for factors.
  • Select Principal Components to Plot
    Specify factors for the X, Y and Z axis in plots.
    • Factor for X Axis
      Specify the factor for the X axis in component plots.
    • Factor for Y Axis
      Specify the factor for the Y axis in component plots.
    • Factor for Z Axis
      Specify the factor for the Z axis in component plots.
  • X Loading Plot
    Specify whether to show the Loading Plot for eigenvectors of X.
  • Y Loading Plot
    Specify whether to show the Loading Plot for eigenvectors of Y.
  • X Score Plot
    Specify whether to show the Score Plot for scores of X.
  • Y Score Plot
    Specify whether to show the Score Plot for scores of Y.
Diagnostics Plot Specify whether to show the diagnotics plots, which includes,
  • Predicted Values v.s. Dependents Plot
  • Residual v.s. Predicted Values Plot
  • Residual vs. Independents Plot
  • P-P plot of the Residuals
Distance Plot Specify whether to show the distance plot
T Square Plot Specify whether to show the t square plot

Output

Specify the destination of output results.

Partial Least Squares Report Specify the sheet for the partial least squares report. The default value is a new sheet in the workbook of input data named as PLSn.
Report Data (Fitted Results) Specify the sheet for the results of the option selected in the Quantities branch. The default value is a new sheet in the workbook of input data named as PLSResultsn.
Report Data (Loadings and X Weights) Specify the sheet for the loading values and the X weights. The default value is a new sheet in the workbook of input data named as PLSLoadingsn.
Report Data (Scores and Residuals) Specify the sheet for the score values and the residual values. The default value is a new sheet in the workbook of input data named as PLSScoresn.