CCDM Procedure
Output Data Tables
PROC CCDM writes the output data tables that you specify in the OUT= option of the OUTPUT and OUTSUM statements. It also produces an output data table of individual severity draws when you specify the OUTDRAW= option. The contents of these output data tables are described in the following sections.
OUTDRAW= Data Table
The OUTDRAW= data table records the individual severity draws.
If you specify the BY statement, then the data table contains variables that you specify in the BY statement. In addition, the OUTDRAW= data table contains the following variables:
_SEVERITYMODEL_
indicates the name of the severity distribution model._COUNTMODEL_
indicates the name of the count model. If you specify the EXTERNALCOUNTS statement, then the value of this variable is "_EXTERNAL_"._DRAWID_
indicates the identifier for the perturbed sample. This variable is created only when you specify the NPERTURBEDSAMPLES= option in the PROC CCDM statement. The value of this variable identifies the perturbed sample. A value of 0 indicates an unperturbed sample._BEGINAGG_
indicates whether the observation starts a new aggregate sample point or continues the previous aggregate sample point. A value of 1 indicates that a new aggregate sample point begins. A value of 0 indicates that the observation continues the aggregate sample point._SEVERITYDRAW_
indicates the value of the severity draw.
OUTSAMPLE= Data Table
The OUTSAMPLE= data table records the full sample of the aggregate loss and aggregate adjusted loss.
If you specify the BY statement, then the data table contains variables that you specify in the BY statement. In addition, the OUTSAMPLE= data table contains the following variables:
_SEVERITYMODEL_
indicates the name of the severity distribution model._COUNTMODEL_
indicates the name of the count model. If you specify the EXTERNALCOUNTS statement, then the value of this variable is "_EXTERNAL_".- <sample variable>
indicates the value of the aggregate unadjusted loss. The name of this variable is the value of the SAMPLEVAR= option in the OUTPUT statement. If you do not specify the SAMPLEVAR= option, then this variable is named_AGGSEV_.- <adjusted sample variable 1> …<adjusted sample variable D>
indicates the values of the D aggregate adjusted losses, where D is the number of adjusted severity symbols that you specify in the ADJUSTEDSEVERITY= option in the PROC CCDM statement. The names of these variables are the values of the ADJSAMPLEVAR= option in the OUTPUT statement, which are applied to the symbols in the order in which you specify them in the ADJUSTEDSEVERITY= option. If you specify fewer names in the ADJSAMPLEVAR= option than symbols in the ADJUSTEDSEVERITY= option, then the variable for each remaining adjusted severity symbol has the same name as the corresponding symbol. If you do not specify the ADJSAMPLEVAR= option in the OUTPUT statement, if you specify only one adjusted severity symbol in the ADJUSTEDSEVERITY= option, and if you specify the ONEADJSEVCOMPAT option, then the adjusted sample variable is named_AGGADJSEV_._DRAWID_-
indicates the identifier for the perturbed sample. This variable is created only when you specify the NPERTURBEDSAMPLES= option in the PROC CCDM statement and the PERTURBOUT option in the OUTPUT statement. The value of this variable identifies the perturbed sample. A value of 0 indicates an unperturbed sample.
Observations that have the same
_DRAWID_variable value are always distributed across all worker nodes for the unperturbed sample (_DRAWID_= 0). For each of the perturbed samples (_DRAWID_
0), observations are either distributed or partitioned across worker nodes, depending on the value of the PERTURBMODE= option. For more information, see the description of the PERTURBMODE=AUTO option.
OUTSUM= Data Table
The OUTSUM= data table records the summary statistics and percentiles of the compound distributions of aggregate loss and aggregate adjusted loss. Only the estimates that you request in the OUTSUM statement are written to the OUTSUM= data table. For more information about the method of naming the variables that correspond to the summary statistics or percentiles, see the description of the OUTSUM statement.
If you specify the BY statement, then the data table contains variables that you specify in the BY statement. In addition, the OUTSUM= data table contains the following variables:
_SEVERITYMODEL_
indicates the name of the severity distribution model._COUNTMODEL_
indicates the name of the count model. If you specify the EXTERNALCOUNTS statement, then the value of this variable is "_EXTERNAL_"._SAMPLEVAR_-
indicates the name of the aggregate loss sample. For an unadjusted sample, the value of the variable is the value of the SAMPLEVAR= option that you specify in the OUTPUT statement. If you do not specify the SAMPLEVAR= option, then the default value is_AGGSEV_.For an adjusted sample, the value of the variable is decided by the value of the ADJSAMPLEVAR= option that you specify in the OUTPUT statement or the name of the adjusted severity symbol. If you do not specify the ADJSAMPLEVAR= option, if you specify only one adjusted severity symbol in the ADJUSTEDSEVERITY= option, and if you specify the ONEADJSEVCOMPAT option, then the value is
_AGGADJSEV_. _DRAWID_indicates the identifier for the perturbed sample. This variable is created only when you specify the NPERTURBEDSAMPLES= option in the PROC CCDM statement. The value of this variable identifies the perturbed sample. A value of 0 indicates an unperturbed sample.