5 ms·
Even if you get the false positive rate low, the low prevalence of criminals versus innocents will lead to a large number of false positives. Let's say one per
by flashman 8y ago
Even if you get the false positive rate low, the low prevalence of criminals versus innocents will lead to a large number of false positives.
Let's say one person in 1000 has an outstanding warrant. Let's say that when the camera sees this person, it has a 99% chance of recognising them correctly. Let's also say that it has only a 1% chance of recognising an innocent person as a suspect.
Under these generous conditions, 10 out of every 11 hits will be false positives. Put another way, if you run the hypothetical system on a football match with 60,000 people in the crowd, you'll find 59 or 60 of the ones with outstanding warrants, and several hundred without.
This is just the same math that was used to show how ineffective PRISM mass surveillance would be: https://bayesianbiologist.com/2013/06/06/how-likely-is-the-nsa-prism-program-to-catch-a-terrorist/ https://bayesianbiologist.com/2013/06/06/how-likely-is-the-n...
- a1369209993 8y agoI think this may be a problem of terminology. When I (and, I think, others) say "low false positive rate" what we're actually talking about is the (more useful for approximation) ratio "fraction of negatives classified positive" over "fraction of population that is positive". You have described a false positive 'rate' of one thousand percent (10 to 1). A genuinely low false positive 'rate' of 1 to 10 would require a actual-rate around 0.01% chance of recognising an innocent person as a suspect. It might help if there was widely known term for 'rate's, rather than using wrong-term-in-scare-quotes. Edit: Also, it's increasingly impossible to have a low false positive 'rate' for one-in-a-million or one-in-a-billion events, and not-technically-lying about how good your detectors are is probably a significant secondary factor in why this sort of thing gets so much flak for "staggering inaccuracy".
- jessaustin 8y agoYou can't get around the fact that rare events happen rarely. False positive rate is the rate of false positives: given that this one event is a positive, what is the chance that it is a false positive? This is important for decision-making in a real situation. Your suggested definition is for a much less helpful statistic, no matter how comforting it might seem to accuse the field of statistics of "not-technically-lying". In fact your idiosyncratic definition has the opposite effect from the one you seem to seek: when "true" negatives are far more common than "true" positives, false positives typically will also be far more common than true positives.
- a1369209993 8y agoYes, I know what a false positive rate is, hence the scare-quotes everywhere I misused "rate". And I'm (obviously) not accusing the field of statistics of not-technically-lying; I'm accusing supporters of face recognition police tools of not-technically-lying. False positive rates are actively misleading when (trying to avoid) answering the question "given that this one event is reported positive, what is the chance that it actually is?".
- flashman 8y agoWell, no, that's not the Wikipedia description of a false positive rate:[1] "The false positive rate is calculated as the ratio between the number of negative events wrongly categorized as positive (false positives) and the total number of actual negative events (regardless of classification)." That ratio in my example is 1 to 99. [1]https://en.wikipedia.org/wiki/False_positive_rate https://en.wikipedia.org/wiki/False_positive_rate