False positive rate (FPR): Difference between revisions

==Introduction==

False positive rate (FPR) is the proportion of actual negative [[example]]s for which the model mistakenly predicted the positive [[class]]. [[Machine learning models]] use [[classification]] [[models]] to predict whether an [[input]] belongs to a certain [[class]] or not. Unfortunately, these predictions aren't always correct and sometimes the model may indicate that an input belongs in one class when it actually doesn't - this is known as a false positive and the rate at which false positives occur is known as the false positive rate (FPR).

==Explain Like I'm 5 (ELI5)==

False positive rate is like when your teacher thinks you did your homework, but you actually didn't. The false positive rate is the number of times your teacher thinks you did your homework when you didn't, divided by the total number of times your teacher thinks you did your homework. A low false positive rate means your teacher doesn't think you did your homework very often when you didn't, which is good. But sometimes it's okay for your teacher to think you did your homework when you didn't, depending on why you didn't do it.