Add Substring Support
At the moment it is easy to include profane words - I can for example include f##kface and it is not picked up as profane. Maybe add the ability to look for specific substrings?
Francois
shared this idea
1 comment
-
Hi Francois,
What might help your requirement is using the "classify = true" parameter with the API to enable machine-learning based text classification feature on the text, in addition to the term-based profanity filtering.For example, if I used your example word in a sentence as in "You f##kface!", or just "f##kface!" you will see a high offensive score ("Category 3") and the ReviewRecommended flag is set to True.
"Classification": {
"ReviewRecommended": true,
"Category1": {
"Score": 0.068794898688793182
},
"Category2": {
"Score": 0.2296631783246994
},
"Category3": {
"Score": 0.76672351360321045
}
},Hope this helps.
