what threshold for the score to accept the prediction ?

facebookresearch / BLINK

Entity Linker solution

MIT License

1.17k stars 231 forks source link

what threshold for the score to accept the prediction ? #37

Open Hadjerkhd opened 4 years ago

Hadjerkhd commented 4 years ago

I'm using BLINK for entity mapping I want to know what is the interval of values for the prediction scores, to be able to fix an acceptable threshold to consider the predicted linking as true. When looking at some example, I've noticed that the values could be positive or negative, and when the returned label from Wikipedia is true, the score is highly positive, is my assumption true ? how we interpret the values of the score ? Thanks in advance

ledw commented 3 years ago

@Hadjer13 Are you using fast mode or not? If not, the prediction scores should be the logit values of the top 10 predictions, so you can apply a softmax on top of it to get probabilities. "when the returned label from Wikipedia is true" : What is this referring to?