The LOESS Procedure
-
Overview
-
Getting Started
-
Syntax
-
Details
Missing Values Output Data Sets Data Scaling Direct versus Interpolated Fitting kd Trees and Blending Local Weighting Iterative Reweighting Specifying the Local Polynomials Smoothing Matrix Model Degrees of Freedom Statistical Inference and Lookup Degrees of Freedom Automatic Smoothing Parameter Selection Sparse and Approximate Degrees of Freedom Computation Scoring Data Sets ODS Table Names ODS Graphics
-
Examples
- References
| SCORE Statement |
- SCORE <DATA=SAS-data-set> <ID=(variable list)> </ options> ;
The fitted loess model is used to score the data in the specified SAS data set. This data set must contain all the regressor variables specified in the MODEL statement. Furthermore, when a BY statement is used, the score data set must also contain all the BY variables sorted in the order of the BY variables. A SCORE statement is optional, and more than one SCORE statement can be used. SCORE statements cannot be used if the DIRECT option is specified in the MODEL statement. The optional ID= (variable list) specifies ID variables to be included in the "Score Results" table.
You find the results of the SCORE statement in the "Score Results" table. This table contains all the data in the data set named in the SCORE statement, including observations with missing values. However, only those observations with nonmissing regressor variables are scored. If no data set is named in the SCORE statement, the data set named in the PROC LOESS statement is scored. You use the PRINT option in the SCORE statement to request that the "Score Results" table be displayed. You can place the "Score Results" table in an output data set by using an ODS OUTPUT statement even if this table is not displayed.
The following options are available in the SCORE statement after a slash (/).
- CLM
requests that
)% confidence limits on the mean predicted value be added to the "Score Results" table. By default the 95% limits are computed; the ALPHA= option in the MODEL statement can be used to change the significance level. The use of this option implicitly selects the model option DFMETHOD=EXACT if the DFMETHOD= option has not been explicitly used. - PRINT <(VAR=variables )>
specifies that the "Score Results" table be displayed. By default only the variables named in the MODEL statement, the variables listed in the ID list in the SCORE statement, and the scored dependent variables are displayed. You can use the VAR= option to specify additional variables in the score data set that are to be included in the displayed output. Note, however, that all columns in the SCORE data set are placed in the SCORE results table, even if you do not request that they be included in the displayed output.
- RESIDUAL | R
requests that residuals be added to the "Score Results" table. If the data set you specify in DATA= option in the SCORE statement does not contain one or more of the model dependent variables, then the corresponding residual values in the "Score Results" table are set to missing.
- SCALEDINDEP
specifies that scaled regressor coordinates be included in the "Score Results" table. This option is ignored if the SCALE= option is not specified in the MODEL statement.
- STEPS
requests that all models evaluated during smoothing parameter value selection be scored, provided that the SELECT= option together with the STEPS modifier is specified in the MODEL statement. By default only the selected model is scored.