CCDM Procedure
Displayed Output
The CCDM procedure optionally produces displayed output by using the Output Delivery System (ODS). You control all output by using the PRINT= option in the PROC CCDM statement. Table 5 relates the PRINT= options to ODS tables.
Table 5: ODS Tables Produced in PROC CCDM
| ODS Table Name | Description | Option |
|---|---|---|
| CompoundInfo | Compound distribution information | Default |
| DataSummary | Input data summary | Default |
| Percentiles | Percentiles of the aggregate loss sample | PRINT=PERCENTILES |
| PerturbedPctlSummary | Perturbation analysis of percentiles | PRINT=PERTURBSUMMARY and NPERTURBEDSAMPLES > 0 |
| PerturbedSummary | Perturbation analysis of summary statistics | PRINT=PERTURBSUMMARY and NPERTURBEDSAMPLES > 0 |
| SummaryStatistics | Summary statistics of the aggregate loss sample | PRINT=SUMMARYSTATISTICS |
| Timing | Timing information for various computational stages of the procedure | PRINT=ALL |
PRINT= Option
This section provides detailed descriptions of the tables that you display by using different PRINT= options.
-
If you do not specify the PRINT= option and if you do not specify the NOPRINT or PRINT=NONE options, then by default PROC CCDM produces the CompoundInfo, DataSummary, and SummaryStatistics ODS tables.
The "Compound Distribution Information" table (ODS name CompoundInfo) displays the information about the severity and count models.
The "Input Data Summary" table (ODS name DataSummary) is displayed when you specify the DATA= data table. The table displays the total number of observations and the valid number of observations in the data table. If you specify the EXTERNALCOUNTS statement, then the table also displays the number of replications and total number of loss events across all replications.
-
If you specify PRINT=PERCENTILES, the "Percentiles" table (ODS name Percentiles) is displayed for the distribution of the aggregate loss. The table contains estimates of all predefined percentiles in addition to the percentiles that you request in the OUTSUM statement.
If you specify the programming statements and the ADJUSTEDSEVERITY= option symbols, then an additional table is displayed for the distribution of each aggregate adjusted loss. These tables also contain estimates of all predefined percentiles in addition to the percentiles that you request in the OUTSUM statement.
-
If you specify PRINT=PERTURBSUMMARY, two tables are displayed for the distribution of the aggregate loss. The "Perturbed Summary Statistics" table (ODS name PerturbedSummary) displays the summary of the effect of perturbing model parameters on the following five summary statistics of the distribution: mean, standard deviation, variance, skewness, and kurtosis. The "Perturbed Percentiles" table (ODS name PerturbedPctlSummary) displays the perturbation summary for all predefined percentiles in addition to the percentiles that you request in the OUTSUM statement.
The tables are displayed only if you specify a value greater than 0 for the NPERTURBEDSAMPLES= option.
If you specify a value of P for the NPERTURBEDSAMPLES= option, then for each summary statistic and percentile, an average and standard error of the set of P values of that summary statistic or percentile are displayed in the respective perturbation summary tables.
If you specify the programming statements and the ADJUSTEDSEVERITY= option symbols, then additional perturbation summary tables are displayed for the distribution of each aggregate adjusted loss.
-
If you specify PRINT=SUMMARYSTATISTICS, the "Summary Statistics" table (ODS name SummaryStatistics) is displayed for the distribution of the aggregate loss. The table contains estimates of the following summary statistics: number of observations in the sample, maximum value in the sample, minimum value in the sample, mean, median, standard deviation, interquartile range, variance, skewness, and kurtosis.
If you specify the programming statements and the ADJUSTEDSEVERITY= option symbols, then an additional table of summary statistics is displayed for the distribution of each aggregate adjusted loss.