Setting Options

Option Name	Description
Methods
Number of components	specifies the number of principal components to compute.
Details
Analyze	specifies whether to compute the principal components from the correlation matrix or the covariance matrix.
Exclude observations with missing values	excludes all observations with missing values from the analysis. However, these observations are not excluded from the output data set. If you do not select this option, the task computes the optimal transformations of the nonmissing values and estimates the missing values that minimize the squared error.
Standardize variance of scores	standardizes the principal component scores in the output data set to mean zero and variance one or a default mean of zero and variance equal to the corresponding eigenvalue.
Specify the maximum number of iterations	specifies the maximum number of iterations.
Plots
By default, these plots appear in the results: a scree plot (eigenvalue by component) an MDPREF plot. This plot is produced with points for each row and vectors for each column. Often, the vectors are short, and a better graphical display is produced when the vectors are stretched. By default, the length of the vectors is 1, so you see the vectors without any stretching. Select the Change length of the MDPREF plot vectors check box to specify the absolute lengths of each vector. Specify the scaling factor in the Vector scaling factor text box. Now, the vector coordinates are all multiplied by n. The relative lengths of the different vectors are important and interpretable, and these lengths are preserved by the stretching. To label the plot points in the graph and to include tooltips, assign variables to the Label MDPREF plot points role. You can also include a variable transformations plot in the results.