I need to be able to select first N most representative points from each cluster calculated by K-means. To do so, I am aiming to calulate the distance of each point to its own cluster center and take the ones with the smallest values. To do this, I have used the K-means transform function:

kmeans= KMeans(n_clusters=18,random_state=2, n_init = 50, max_iter=500)
clusterer= kmeans.fit(data)
distances= kmeans.transform(data)[:,0]

To validate the results, I have calculated the points closest to the cluster centers:

closest_org, _ = pairwise_distances_argmin_min(clusterer.cluster_centers_, data)

and the points that I have calculated as the closest ones are not the one that have a smallest distance. Could you please advise on the optimal and correct way to calculate get the N closest points from each cluster?


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.