Limit text found in Entity Extraction model

[cortese98]

I have a model that is searching for a Product in email. The model is finding the product at a high percentage, it is also recognizing other text as Products though. What is the best way to train the model to not recognize other text as that Product tag? 

[CedrickB]

Hi,

You could try creating a custom entity extraction model removing all default entities by

- Clicking on the wheel

- Then clicking on "Clear all" and apply

Then you will create a new custom entity "MyProduct" and will tag it in the training data set

Let me know if this works for you.

[cortese98]

I did this already. The issue is I am getting additional items tagged to my custom entity. Do I just need to add additional example sentences? 

Example below: 

Products(Value: Fiber) accurate. 

Products(Value: 7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0)Products(Value: lR9JZbxFLAxGMSOtgtIs6AgMtNF9rHpBPBpNwAXdpx8) Not accurate