Computer scienceData scienceMachine learningClustering

K-Means

k-means++

Report a typo

Imagine you work with patient data from our HyperClinic. We'd like to cluster our patients into groups, so we may gain new insights about them. You find the random initialization of k-means too naive, so you'd like to try out k-means++. You've selected a data point for the first centroid. Which data point should be selected for the second centroid initialization? Below you find data points together with their distances to the first centroid.

Select one option from the list

X_2, d=55.5

X_3, d=2.32

X_1, d=52.5

X_4, d=22.5

___

Create a free account to access the full topic