Am trying to figure out how we do the Quality calculation on a list of Rules in this algorithm:
I understood how to calculate the Quality on a Single Rule, using this formula:
But when it comes to the case of a list of Rules, how do we find out the Quality of the list of Rules?
We will find the Quality of Rules one by one and sums it up?
EDIT
Based on this paper, the Quality of the list of Rules is found using this formula:
So almost getting my query solved with this. But am having doubt regarding the numerator of it. I mean the total examples
is the total number of rows in the validation set isn't it? And the number of correct classification
means number of rows in the validation set that matches the Rules of our list? I mean those rows that matches the Rule Terms as well as the Rule's class value?