This algorithm randomly selects k data in the dataset D as the initial cluster center point.
(b)
This algorithm finds distance from all remaining data points to the k cluster center points and selects the closest point to itself as its cluster center point.
(c)
The algorithm calculates the center points of all classes in turn and uses these center points as the new cluster centers.
(d)
The algorithm iterates steps b and c repeatedly until the error average and value are less than the specified threshold. The algorithm ends.
