The basic step of k-means clustering is simple. In the beginning, we determine number of cluster K and we
assume the centroid or center of these clusters. We can take any random objects as the initial centroids or the
first K objects can also serve as the initial centroids.