KMeans Quiz Q.10

#1

How can the answer be both A and B when B is the graph for batch gradient descent and in Kmeans we apply coordinate descent.

#2

hey @devanshmarwaha27 ,
we really apologize for such late response.

your doubt is mainly about where do we perform gradient descent in K means…correct?
So the point is there are many drawbacks of K Means with coordinate ascent

  1. This algorithm doesn’t performs well if the features are correlated to each other.
  2. Scaling of features has a high effect on its working
  3. Taking very much time on large datasets

Just to over come these drawbacks to an extent , stochastic gradient descent is used in K means. You read about in detail on this link. Major use of gradient descent in K means is to make it converge faster on very large datasets.

I hope this would have resolved your doubt.
Thank You and Happy learning :slightly_smiling_face:.