Right now, we just make epsilon calculated automatically at each training and compare it directly with that value.
We should add a SlideBar in the Settings where the user can control the multiple of epsilon that the prediction algorithm compared with (i.e. if predicted < epsilon, user is impostor) from 0.3 to 30x epsilon (slidebar should scale logarithmically). Default should be 1x
This allows the user to control the strictness of the lock screen to what they like.
Right now, we just make epsilon calculated automatically at each training and compare it directly with that value.
We should add a SlideBar in the Settings where the user can control the multiple of epsilon that the prediction algorithm compared with (i.e. if predicted < epsilon, user is impostor) from 0.3 to 30x epsilon (slidebar should scale logarithmically). Default should be 1x
This allows the user to control the strictness of the lock screen to what they like.