Talk:crisp

"Crisp" in logic and machine learning

Here is another meaning of the adjective "crisp" in English, but I'm not adding it since I don't have references:

• "Crisp" logic is the opposite of "fuzzy" logic, i.e., in crisp logic allows only true and false, and no values in between. This is as opposed to fuzzy logic which also allows degrees of "true" and "false".

The term is also used with classification in machine learning. I.e., "crisp classification" refers to a classification method that assigns exactly one class to an input, as opposed to probabilistic classification, which assigns probabilities to the input being in each class.

Synonym: "boolean"

Antonyms: "fuzzy", "probabilistic"

Cheers –Jérôme (talk) 14:03, 13 November 2015 (UTC)