cluster analysis - Interpreting the results of R Mclust package -
i'm using r package mclust estimate number of clusters in data , result:
clustering table: 2 7 8 9 205693 4465 2418 91 warning messages: 1: in map(z) : no assignment 1,3,4,5,6 2: in map(z) : no assignment 1,3,4,5,6
i have 9 clusters best, has no assignment 5 of clusters. mean want use 9 or 5 clusters?
if answer can found somewhere online, link appreciated. in advance.
most likely, method did not work @ on data...
you may try other seeds, because when "lose" clusters (i.e. become empty) means seeds not chosen enough. , cluster 9 pretty gone, too.
however, if data generated mixture of gaussians, it's hard find such bad starting point... likely, of results bad, because data not satisfy assumptions.
judging cluster sizes, i'd have 1 cluster , lot of noise...
have visualized , validated results?
don't blindly follow number. validate.
Comments
Post a Comment