Generate Aggregation (Advanced)
Synopsis
This operator allows you aggregate a selection of attributes per example.
Description
This operator works very similar to the Generate Aggregation operator. Both operators allow you aggregate per example (i.e. row based). You can select the aggregation function using the 'function' parameter. Each function may have one or more parameters. The name of the generated attribute is always the function name you selected.
The differentiation to Generate Aggregation is, that this operator allows you more complex aggregation functions.
The differentiation to Aggregate is, that this operator allows you to aggregate row-wise, while aggregate is performing aggregations on a columnar basis.
Input
exa
The ExampleSet you want to perform the aggregation on.
Output
exa
The ExampleSet with aggregations
ori
The original ExampleSet.
Parameters
Function
This parameter allows you to set the aggregation function. Note that the name of the generated attribute is always the function name
- kth min: Calculates the kth minimum of a set of values. Example: If the input data is 30,50,10,2 and you select k=2, you get the 2nd smallest value; In this case, 10.
- kth max: Calculates the kth maximum of a set of values. Example: If the input data is 30,50,10,2 and you select k=2, you get the 2nd biggest value; In this case, 30.
K
Configuration for 'kth-min' and 'kth-max' aggregation function. Used to define which minimum to extract.
Add column name
If set to true an additional column "Column Name" is added, which contains the name of the column with the kth-min/max value. In case of multiple values only the first appearance is added.