4 ms·
> A possible solution to this is creating data points for everyone A better solution is to use better statistics. If you think you need to collect redundant e
by jlokier 6y ago
> A possible solution to this is creating data points for everyone
A better solution is to use better statistics.
If you think you need to collect redundant extra data about innocents just to improve inferences using the data you have, you're doing statistics wrong.
This has come up with DNA databases too.
Prosecutor's Fallacy:
https://en.wikipedia.org/wiki/Prosecutor%27s_fallacy https://en.wikipedia.org/wiki/Prosecutor%27s_fallacy
https://www.cebm.ox.ac.uk/news/views/the-prosecutors-fallacy https://www.cebm.ox.ac.uk/news/views/the-prosecutors-fallacy
"The Prosecutor’s Fallacy is most often associated with miscarriages of justice. It’s when the probability of innocence given the evidence is wrongly assumed to equal an infinitesimally small probability that that evidence would occur if the defendant was innocent. Consequently, highly improbable innocent explanations have led to the assumption of guilt,"
- lxnn 6y agoIt's essentially when information that was used to identify the suspect is presented as confirmation of their guilt. > The witness described a tall, white, black-haired man with a tattoo on his right ankle. Only .001% of the population match this description, so there is only a .001% chance that the suspect is innocent. Of course, this ignores the fact that they compared the description against a database of one million people, among whom you would expect to find 10 coincidental matches. For the same reason more data is required to select a model than to validate a model in machine learning.