monolithMon, Apr 7, 2008 This is a bit technical issue, but here's an plastic example :
Later, some kind of admin comes along and just clicks off the things that really happened (presented as yellow boxes). That is if thing happened - the end result is 100% and if not - 0%. Now, if my call was 80% and end result is 100% you may say that my accuracy was 80%, and if end result is 0% we shall say that it was 20%. Ok?
As you can see from this picture above, we can allow a multitude of things of the same time-rank to happen, as not all of predicted things are exclusive with each other. And also the sum within one time-rank doesn't have to be 100%, which is an error from my previous pictures.
This is a bit technical issue, but here's an plastic example :
Later, some kind of admin comes along and just clicks off the things that really happened (presented as yellow boxes). That is if thing happened - the end result is 100% and if not - 0%. Now, if my call was 80% and end result is 100% you may say that my accuracy was 80%, and if end result is 0% we shall say that it was 20%. Ok?
As you can see from this picture above, we can allow a multitude of things of the same time-rank to happen, as not all of predicted things are exclusive with each other. And also the sum within one time-rank doesn't have to be 100%, which is an error from my previous pictures.