0
$\begingroup$

no. of cluster = 10, data points = 6000

library(ppclust)
cm <- fcm(x,centers = cen)

takes ~ 10 minutes

library(e1071)
cm <- cmeans(x,cen,1000)

takes ~ 1 minute

the only reason I prefer ppclust is because it allows nstart, meaning I am not stuck with local optima.

Given both are doing same thing, why is one taking 10x time then other?

$\endgroup$
0
$\begingroup$

Look at the source code. The guts of e1071::cmeans are written in C, but ppclust::fcm looks like all R and has some ugly looking eval stuff which might make it dead slow.

$\endgroup$

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.