How do I pre-process new text before the Score Node in Text Mining?

[MDH]

Hi all

Appreciate this in advance.

I have a nicely performing Text Rule Builder model in EM14.1.

When I present new text data to the Score node, I know I must declare the data Role = Score.

But this means I can't put a Parsing Node between the new data and the Score node, because Parsing requires Train or Raw Role.

How do I get around this? Is it OK to not pre-process the new data to be scored?

Best

MDH

[CraigDeVault · Accepted solution]

The Score code contained in the Score node will contain all of the information needed to score a new set of textual data.  You do not need to run a Text Parsing node between the score data set and the Score node.  The score code will apply all of the parsing from the original set of observations to the new data set of observations that is to be scored.

[MDH]

Thanks so much, that's good to know!