How to "clean terms" in Contrast Analysis?

CorText Manager Q&A forum › Category: Data exploration › How to "clean terms" in Contrast Analysis?
0 Vote Up Vote Down
Rome asked 5 years ago

Hi everyone, 
i’m trying to run a Contrast Analysis opposing two targets, but I get many “useless terms” as “http” “les” etc. 
 
Is there any way to get rid of them, they parasite the other words.
On a previous post, someone kinda ask the same question, and she’s been told to use “categorical” instead of “text” but it still doesnt work..
 
Any ideas ?
 
Thanks a lot

Jean-Philippe Cointet Staff replied 5 years ago

Hi Rome,
As long as your field is made of textual content, you should proabably stick to “text” and avoid “categorical in my opinion.
You may try to use the alternative tokenizer that is slower but may be more efficient (it may not cut URLs and better manage accents or other characters).