Algorithm 2: Midpoint method for initial CH nomination

Input:

X = consists of a total n number of data points.

Copt = optimal cluster density

Output:

initial centroids of the Copt clusters.

Steps:

1: Origin ($x_o,y_0$), Data point i ($x_i,y_i$)

For i = 1:n

$l\left(i\right)=\sqrt{{\left(x_o-x_i\right)}^2+{\left(y_0-y_i\right)}^2}$

end

2: Sort ($l$)

3: n/Copt.

4: The middle point value of each set is considered as the initial centroid.