Quote:
Originally Posted by nbritton
Who? 80% accurate, I think that is actually quite good.
|
http://www.behaviorismandmentalhealt...et-al-2010.pdf
I think the issue is false positives...only 83% of controls were correctly identified so there is a 17% chance they would tell you you had sz when you didn't...even pdocs are better than that. It's also testing a pattern of some 51 markers so it's not any one thing...but it is interesting.