Hi,
I am running Text IQ on freetext responses (in Dutch).
I notice that quite often, the sentiment score calculated by Qualtrics is not correct.
In the edit mode, you can edit the sentiment scores of the post overall and also the topic sentiment.
However, does Text IQ learn from that? Does it then know that other similar posts should also get another score? (if you manually edit enough responses of course, just changing 1 entry is not enough for any model to learn from it).
Idem when manually assigning topics to a post, can that cause other similar posts to also be tagged with that topic?
Best answer by JenCX
View original