During the two classification process, we can set any category as a positive example.

Question

Options

• ATRUE
• BFALSE

• A
  84% (47)
• B
  16% (9)

Explanation

In binary (two-class) classification, the labeling of which class is "positive" and which is "negative" is a human convention, not a mathematical requirement. You can freely designate either class as positive depending on your context - for instance, in email filtering you might call spam "positive," but you could just as validly flip it. The choice affects how metrics like precision and recall are reported, but it does not change the underlying model or data.

Why B is wrong: There is no rule - statistical, algorithmic, or otherwise - that restricts which class must serve as the positive example. The classification algorithm operates symmetrically on the two classes; the "positive" label is purely a naming convention applied afterward.

Memory tip: Think of a coin flip - heads vs. tails. You can call either side "heads" and the coin doesn't care. Similarly, in binary classification, either class can be your "positive" - the model just learns the boundary between the two, regardless of which label you assign.

No community discussion yet for this question.