Aggregation (Extended)

Use the Expression group of parameters:

• Dimension. The dimension, in which the aggregation is executed. The drop-down list includes all database dimensions, except for the calendar dimension, measurement units, revisions, and mnemonics.

• Value. Aggregation expression. Attribute names are specified in figured brackets, for example: {BCI}+{BCA}. All the available aggregation expressions are shown in the menu that opens on pressing CTRL + SPACE.

• Filter. Limitation on rows, on which aggregation is executed. Work with the box in the same way as with the Output Variable box on the Parameters panel.

Use the following parameters:

• Method. Select one of the available aggregation methods in the drop-down list.

• Weights. Set an expression for aggregation weight. The drop-down list is available if a method of weighted aggregation (for example, weighted sum) is selected. The list contains elements of the dimension selected in the Dimension box. Elements are not selected by default.

• Ignore Series with Empty Weights. The checkbox determines whether the series with empty weights are ignored. If the series are ignored, aggregation calculation takes less time. The checkbox is deselected by default and is available if a weighted aggregation method is selected (for example, weighted sum).

• Level. The checkbox determines whether the Level function that smoothes obtained data is used on calculating the aggregation.

The checkbox is available when the following methods or their "weighted" analogs are selected:

1. • Weights.

   • Weighted Geometric Mean.

   • Number of Observations.

   • Number of Missing Values.

   • Sum.

   • Average.

   • Median.

   • Mode.

   • Percentile.

   • Composition Relevance.

The checkbox is selected automatically and dimmed when the following methods or their weighted analogs are selected:

1. • Percentage Change Average.

   • Log Difference Average.

   • Percentage Change Median.

To apply the Rebase function allowing to rebase data from one frequency to another, select the Rebase checkbox.

NOTE. The Rebase checkbox is available in the same cases as the Level checkbox that sets aggregation calculation method parameters.

Set additional rebase parameters:

• Frequency. Calendar frequency that is used to calculate the Rebase function. It is available to change if model calculation frequency differs from annual frequency.
  The Rebase function can be used for annual frequency or for lower frequencies but not lower than the model calculation frequency. For example, a model calculated by quarterly frequency. Thus, only the following options are available in the Frequency drop-down list: Annual, Semi-annual and Quarterly.

• Base Period. If the radio button is selected, the box determines the period when the Rebase function is calculated. The format of setting a period depends on the selected calculation frequency. For example, if the annual frequency is selected, only year should be specified; if the semi-annual frequency is selected, half-year and year should be specified, and so on. If the Frequency box is unavailable in the Rebase group, the period format depends on the Frequency box of the Calculation Periods panel.

• Parameter. If the radio button is selected, the period of calculating the Rebase function is determined via the selected parameter of the metamodel containing the given model in its calculation chain. The list contains a list of parameters referring to the calendar dictionary. The Parameter radio button is available if corresponding parameters are available for the model.

To set aggregation relevance with taking into account missing data, use the Composition Relevance group of parameters:

• Threshold. Threshold value for calculation of aggregation relevance. Available values range: [0; 100]. If threshold value is 0, aggregation is calculated without taking into account missing data. If threshold value is 100, every missing data is taken into account on calculating aggregation. The value that is greater than zero and less than hundred means that the missing data is considered according to the selected threshold.

• Comparator. Comparator is used to estimate aggregation relevance. It contains a list of elements for the dimension selected in the Dimension box on setting up aggregation expression.

Relevance is calculated as follows: if the sum of values of the Comparator element, divided by the sum of all values is less than the Threshold value, the aggregation is not calculated.

Total period is a period, on which all aggregated series have data. To calculate aggregation in the total data period, select the Use of Total Period checkbox and set the following parameter:

• Minimum Number of Observations. The checkbox determines whether the minimum number of observations is limited in the total period. The checkbox is deselected by default, and there is no limit.
  If the checkbox is selected, the minimum number of observations must be specified in the corresponding value editor. By default, the total period must contain at least three observations.

The Use of Total Period checkbox is deselected by default, and aggregation is calculated on the model period.

To calculate aggregation considering data hierarchy, select the Hierarchy Consideration checkbox and set the following parameters:

• Entire Hierarchy. It is used by default. Aggregation is calculated by all hierarchy levels.

• First Level Only. Aggregation is calculated only one level down the hierarchy.

• Consider Source Element. The checkbox is deselected by default, and the source element is not considered in aggregation. If the checkbox is selected, the source element is considered in aggregation.

NOTE. All hierarchical dimensions are used on calculating the aggregation considering hierarchy.

The Hierarchy Consideration checkbox is deselected by default, and calculation is executed without considering hierarchy.