Algorithm 1 Oversampling adjustment algorithm.

Input: Adaptive sampling: $AS$, Array of features of each block: $F$, Block size: $N_B$, Number of blocks: $TB$ Output: Nonoversampled adaptive sampling: $A{S}^{NOA}$ 1: for each AS 2: if the sampling in AS is larger than $N_B$ 3: OS ← 0    //This variable is used to store the amount of oversampling among all blocks 4: NOSI [] ← empty //This variable stores the indices of nonoversampled blocks 5: for counter←1 to TB 6:  if$AS\left(counter\right)>N_B$ 7:    $OSVal\leftarrow AS\left(counter\right)-N_B$ 8:    $OS \leftarrow OS + OSVal$ 9:    $AS\left(counter\right)\leftarrow AS\left(counter\right)-OSVal$ //Subtract the excess sampling from oversampled blocks 10:  Else 11:   NOSI [] ← counter    //Add counter to the array NOSI 12:  end if 13: end for 14: NOSF ← 0       //This variable stores the total feature values of nonoversampled blocks 15: for counter←1 to count(NOSI) 16:   $NOSF\leftarrow NOSF+F\left(NOSI\left(counter\right)\right)$ 17   end for 18:   for counter ←1 to count(NOSI) 19:     $AS\left(NOSI\left(counter\right)\right)\leftarrow AS\left(NOSI\left(counter\right)\right)+\left(\left(\frac{F\left(NOSI\left(counter\right)\right)}{NOSF}\right)\times NOSF\right)$ 20:   end for 21: end if 22: end for 23:$A{S}^{NOA}\leftarrow AS$