1
$\begingroup$

I am analyzing implementation of K-means clustering algorithm in MadLib project. Here K-means algorithm uses Canopy clustering for initial set of Centroid.I am just wondering , are there any other clustering algorithm, that can be used at place of Canopy for better performance?
Below is the link for details description of Canopy Clustering.
canopy clustering

$\endgroup$
3
  • $\begingroup$ There are several k-means algorithms. Can you define the Canopy algorithm you are referring to & what's wrong with it? $\endgroup$ Jan 10, 2016 at 13:31
  • $\begingroup$ there is nothing wrong with it, i am just wants to make it more efficient and fast. Here is the link for canopy algorithm. $\endgroup$
    – Sherlock
    Jan 10, 2016 at 14:09
  • $\begingroup$ i have updated the question with link. $\endgroup$
    – Sherlock
    Jan 10, 2016 at 14:15

1 Answer 1

1
$\begingroup$

A popular initialization for k-means is k-means++.

It chooses initial seeds randomly, weighted by their distance from the previous choices. I.e. it prefers objects that are further apart.

$\endgroup$

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.