In here Andrew mentioned the threshold for predict 1 = 0.3, then in order to predict 0 the threshold is 0.3 as well,
I am not quit understand, isn’t if the threshold for predicting 1 equals to 0.3, then predicting y == 0 should equal to 0.7?
In here Andrew mentioned the threshold for predict 1 = 0.3, then in order to predict 0 the threshold is 0.3 as well,
@Bio_J so I did my ML studies elsewhere, but for one thing I should note is that in other contexts you may hear ‘recall’ labeled as ‘sensitivity’ (same thing-- but that is how I learned it).
Also note he points out the relationship between precision and recall is not a linear one.
But I believe what he is saying here is not ‘predicting 1 == 0.3’, he is saying any value between 0 and anything greater than (also equals) 0.3, predict 1. Any number above (importantly not including) 0.3 to 1, predict 0.
He is giving you a range, not a specific probability value.
Thx, missed the equal sign in the pic, now it makes sense