
305Model Development and Interpretation
MARE
P
i
i
=
−
=
∑
1
1
N
A
A
i
i
where:
Nrefers to the total number of instances in a given data set
P
i
refers to the predicted value of an instance i
A
i
refers to the actual value of an instance i
7.6.3 PRED (A)
It is calculated from the relative error. It is dened as the ratio of the instances that have
an error value (absolute relative error [ARE]) less than or equal to “A” error divided by the
total number of instances in the data set. Given the lowest value of A, the higher the value
of Pred (A), the better it is as the prediction model. Generally, the value of “A” that is used
by most of the studies is 25%, 50 ...