I'm trying to use k-means for segmentation but would like to specify the centroid myself rather than letting k-means select a random seed point. What are the different ways in which the centroid of a cluster can be realized rather than letting k-means find the centroid ?
No products are associated with this question.
0 Comments