Inertia simply means the sum of the squared distances from sample points to their cluster centroid. The smaller the inertia, the closer the cluster members are to their cluster centroid (that's also what KMeans optimizes when choosing centroids). In this context, the elbow method may be helpful (https://bl.ocks.org/rpgove/raw/0060ff3b656618e9136b/9aee23cc799d154520572b30443284525dbfcac5/)
Maybe also take a look at the silhouette metric for choosing K: http://scikit-learn.org/stable/auto_examples/cluster/plot_kmeans_silhouette_analysis.html Best, Sebastian > On Feb 20, 2018, at 5:14 PM, Shiheng Duan <[email protected]> wrote: > > Yes, but what is used to decide the optimal output? I saw on the document, it > is the best output in terms of inertia. What does that mean? > Thanks. > > On Wed, Feb 14, 2018 at 7:46 PM, Joel Nothman <[email protected]> wrote: > you can repeatedly use n_init=1? > > _______________________________________________ > scikit-learn mailing list > [email protected] > https://mail.python.org/mailman/listinfo/scikit-learn > > > _______________________________________________ > scikit-learn mailing list > [email protected] > https://mail.python.org/mailman/listinfo/scikit-learn _______________________________________________ scikit-learn mailing list [email protected] https://mail.python.org/mailman/listinfo/scikit-learn
