Having an "id" token when textmining

kochankochan Member Posts: 11 Contributor II

I would like to do textmining on a dataset with texts the contains the word "id" among others. In my training process I save a wordlist, which I can see contains the token "id". When I later want to classify new data with the model I have build and therefore want to use the wordlist I get a warning, because for example the SingleTextInput operator already has created an "id" attribute.

I can get around the problem by editting the wordlist file and changing the "id" token to "id_", but then of course "id" token will not contribute to the classification.

Is there a way around this problem?

Regards,

Andreas
Sign In or Register to comment.