Adding language variation
The text produced by an NLG system will always vary depending on the values in the input data. This is called data-driven variation. The same data will always produce the same output. Conversely, a different data set (or changes in the same data set), produces different output from the system.
A way to make your text more interesting is adding random variation. This means the same input data will produce several different outputs.
When adding variation, be aware that testing will become more complicated. See 4 — Testing NLG Studio projects for more details.
Tip
When adding variation in scripts, be careful to retain the original, intended meaning; don’t let the meaning change as you change the wording used to express it.
In Arria NLG Studio, the chooseAtRandom function makes it easy to add language variation. In the screenshot below, you can see the chooseAtRandom function is used to vary the sentence describing overall revenue in the Oil & Gas segment.
The remainder of this section is organized as follows:
