Talk:Data Mining Algorithms In R/Clustering/K-Means

Two things in particular seem to be missing:

1. When I plotted the country example as shown, I only got a hello single plot, not a grid of plots. What command makes the grid?

2. How can you clearly label which clusters in the plot go with which categories? Color is used in the country example but you need some kind of key to show which color goes with which group of countries.

[unsigned post]

-

The description of the different algorithms available in R's kmeans function seems inadequate. ?kmeans clearly shows that Forgy and Lloyd options are equivalent. Copied directly from kmeans documentation: "algorithm: character: may be abbreviated. Note that ‘"Lloyd"’ and ‘"Forgy"’ and alternative names for one algorithm."

The typo of "and" instead of "are" is unfortunate, but this line from the kmeans code should clarify things: nmeth <- switch(match.arg(algorithm), `Hartigan-Wong` = 1,        Lloyd = 2, Forgy = 2, MacQueen = 3)

In general, the "four" (three) algorithms are described very generically in a way that does not illustrate the (relatively subtle) differences between them. 142.105.168.230 (discuss) 04:27, 5 December 2013 (UTC)