The OUTPUT statement creates a data table that contains observationwise statistics that PROC PHSELECT computes after fitting the model. In order to avoid data duplication for large data tables, the variables in the input data table are not included in the output data table unless you specify them in the COPYVAR= option.
The output statistics are computed based on the final parameter estimates. If the optimization does not converge, then the output data table is not created.
For observations in which only the censoring variable is missing, values of the linear predictor and the predicted survival probabilities are computed even though these observations do not affect the model fit. This enables, for example, predicted survival probabilities to be computed for new observations.
You must specify the following option:
OUT=CAS-libref.data-table
names the output data table for PROC PHSELECT to use. You must specify this option before any other options. CAS-libref.data-table is a two-level name, where
CAS-libref
refers to a collection of information that is defined in the LIBNAME statement and includes the caslib, which includes a path to where the data table is to be stored, and a session identifier, which defaults to the active session but which can be explicitly defined in the LIBNAME statement. For more information about CAS-libref, see the section Using CAS Sessions and CAS Engine Librefs.
data-table
specifies the name of the output data table.
You can also specify the following syntax elements:
COPYVAR=variable
COPYVARS=(variables)
transfers one or more variables from the input data table to the output data table.
keyword <=name>
specifies a statistic to include in the output data table. Four statistics (DFBETA, RESSCH, RESSCO, and WTRESSCH) are multidimensional; each has a dimension equal to the number of parameters in the model. For the one-dimensional statistic, you can name the variable name. For the multidimensional statistic, name is used as the prefix; for example, if X1–X10 are the explanatory variables in the model, specifying RESSCH=sch creates the variables schX1–schX10 as the 10-dimensional Schoenfeld residual variables. If you do not provide a name, the PHSELECT procedure assigns a default name based on the type of statistic that is requested.
Table 16.7 summarizes the keywords available in the OUTPUT statement.
specifies the predicted cumulative hazard function at the observed time. The default name is _CUMHAZ_.
DFBETA
requests the approximate changes in the parameter estimates when the jth observation is omitted. These variables are a weighted transform of the score residual variables and are useful in assessing local influence. The default name that is used as a prefix is _DFBETA_.
LD
specifies the approximate likelihood displacement when the observation is omitted. This diagnostic can be used to assess the impact of each observation on the overall fit of the model. The default name is _LD_.
RESDEV
specifies the deviance residual. This variable is a transform of RESMART to achieve a more symmetric distribution. The default name is _RESDEV_.
RESMART
specifies the martingale residual, which, at the observed time t, can be interpreted as the difference over in the observed number of events minus the expected number of events. The default name is _RESMART_.
RESSCH
requests the Schoenfeld residuals, which are useful in assessing the proportional hazards assumption. Schoenfeld residuals are computed only at uncensored times and are missing for censored times. If an effect in the MODEL statement is not included in the final model, the corresponding Schoenfeld residuals are set to missing. The default name that is used as a prefix is _RESSCH_.
RESSCO
requests the score residuals, which are a decomposition of the first partial derivative of the log likelihood. They can be used to assess the leverage that each subject exerts in the parameter estimation. The default name that is used as a prefix is _RESSCO_.
ROLE
specifies the numeric variable that indicates the role that each observation plays in fitting the model. The default name is _ROLE_. Table 16.8 shows how this variable is interpreted for each observation.
Table 16.8: Role Interpretation
Value
Observation Role
0
Not used
1
Training
2
Validation
3
Testing
If you do not partition the input data by specifying a PARTITION statement, then the role variable value is 1 for observations that are used in fitting the model and 0 for observations that have at least one missing or invalid value for the response, regressor, frequency, or weight variables.
STDXBETA
specifies the standard error estimates of XBETA. The default name is _STDXBETA_.
SURVIVAL
requests the predicted survival probabilities at the observed times. The default name is _SURVIVAL_.
WTRESSCH
requests the weighted Schoenfeld residuals, which are useful in investigating the nature of nonproportionality if the proportional hazard assumption does not hold. If an effect in the MODEL statement is not included in the final model, the corresponding weighted Schoenfeld residuals are set to missing. The default name that is used as a prefix is _WTRESSCH_.
XBETA
specifies the linear predictor. The default name is _XBETA_.