The TRANSFORM statement lists the variables to be analyzed (variables) and specifies the transformation (transform) to apply to each variable listed. You must specify a transformation for each variable list in the TRANSFORM statement. The variables are variables in the data set. The toptions are transformation options that provide details for the transformation; these depend on the transform chosen. The toptions are listed after a slash in the parentheses that enclose the variables.
For example, the following statements find a quadratic polynomial transformation of all variables in the data set:
proc prinqual; transform spline(_all_ / degree=2); run;
Or, if N1
through N10
are nominal variables and M1
through M10
are ordinal variables, you can use the following statements:
proc prinqual; transform opscore(N1N10) monotone(M1M10); run;
The following sections describe the transformations available (specified with transform) and the options available for some of the transformations (specified with toptions).
There are three types of transformation families: nonoptimal, optimal, and other. The families are described as follows:
preprocess the specified variables, replacing each one with a single new nonoptimal, nonlinear transformation.
replace the specified variables with new, iteratively derived optimal transformation variables that fit the specified model better than the original variable (except in contrived cases where the transformation fits the model exactly as well as the original variable).
are the IDENTITY and SSPLINE transformations. These do not fit into either of the preceding categories.
Table 74.2 summarizes the transformations in each family.
Table 74.2: Transformation Families
Transformation 
Description 

Nonoptimal Transformations 

Inverse trigonometric sine 

Exponential 

Logarithm 

Logit 

Raises variables to specified power 

Transforms to ranks 

Optimal Transformations 

Linear 

Monotonic, ties preserved 

Monotonic Bspline 

Optimal scoring 

Bspline 

Monotonic, ties not preserved 

Other Transformations 

Identity, no transformation 

Iterative smoothing spline 
The transform is followed by a variable (or list of variables) enclosed in parentheses. Optionally, depending on the transform, the parentheses can also contain toptions, which follow the variables and a slash. For example, the following statement computes the LOG transformation of X
and Y
:
transform log(X Y);
A more complex example follows:
transform spline(Y / nknots=2) log(X1 X2 X3);
The preceding statement uses the SPLINE transformation of the variable Y
and the LOG transformation of the variables X1
, X2
, and X3
. In addition, it uses the NKNOTS= option with the SPLINE transformation and specifies two knots.
The rest of this section provides syntax details for members of the three families of transformations. The toptions are discussed in the section Transformation Options (toptions).
Nonoptimal transformations are computed before the iterative algorithm begins. Nonoptimal transformations create a single new transformed variable that replaces the original variable. The new variable is not transformed by the subsequent iterative algorithms (except for a possible linear transformation and missing value estimation).
The following list provides syntax and details for nonoptimal variable transformations.
Optimal transformations are iteratively derived. Missing values for these types of variables can be optimally estimated (see the section Missing Values). See the sections OPSCORE, MONOTONE, UNTIE, and LINEAR Transformations and SPLINE and MSPLINE Transformations in Chapter 97: The TRANSREG Procedure, for more information about the optimal transformations.
The following list provides syntax and details for optimal transformations.
If you use a nonoptimal, optimal, or other transformation, you can use toptions, which specify additional details of the transformation. The toptions are specified within the parentheses that enclose variables and are listed after a slash. For example:
proc prinqual; transform spline(X Y / nknots=3); run;
The preceding statements find an optimal variable transformation (SPLINE) of the variables X
and Y
and use a toption to specify the number of knots (NKNOTS=). The following is a more complex example:
proc prinqual; transform spline(Y / nknots=3) spline(X1 X2 / nknots=6); run;
These statements use the SPLINE transformation for all three variables and use toptions as well; the NKNOTS= option specifies the number of knots for the spline.
The following sections discuss the toptions available for nonoptimal, optimal, and other transformations.
Table 74.3 summarizes the toptions.
Table 74.3: Transformation Options
Option 
Description 

Nonoptimal Transformation 

Uses original mean and variance 

Parameter Specification 

Specifies miscellaneous parameters 

Specifies smoothing parameter 

Spline 

Specifies the degree of the spline 

Spaces the knots evenly 

Specifies the interior knots or break points 

Creates n knots 

Other toptions 

Renames variables 

Reflects the variable around the mean 

Specifies transformation standardization 