Figure 16.7 The outcome of clustering in K-means depends on the initial seeds.
For seeds $d_2$ and $d_5$, K-means converges to {\{$d_1, d_2, d_3\}, \{$d_4, d_5, d_6\}$}, a suboptimal
clustering. For seeds $d_2$ and $d_3$, it converges to {\{$d_1, d_2, d_4, d_5\}, \{$d_3, d_6\}$}, the global
optimum for K = 2.