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
| Scoring Data Sets |
One or more SCORE statements can be used with PROC LOESS. A data set that includes all the variables specified in the MODEL and BY statements must be specified in each SCORE statement. Score results are placed in the ScoreResults table. This table is not displayed by default, but specifying the PRINT option in the SCORE statement produces the table. If you specify the CLM option in the SCORE statement, confidence intervals are included in the ScoreResults table.
Note that scoring is not supported when the DIRECT option is specified in the MODEL statement. Scoring at a point specified in a score data set is done by first finding the cell in the kd tree containing this point and then interpolating the scored value from the predicted values at the vertices of this cell. This methodology precludes scoring any points that are not contained in the box that surrounds the data used in fitting the loess model.