compare the results of the single run to the 2,500 runs.For the two results, compute the sum of squared residuals given by
$$\text{SSR} = \sum_{g = 1}^G \sum_{i: \mathbf{x}_i \in C_g}d(\mathbf{x}_i,\bar{\mathbf{c}}_g)^2$$
where \(G\)
is the number of groups, \(C_g\)
is the set of points in cluster \(g\)
, and \(d(\mathbf{x}_i,\bar{\mathbf{c}}_g)\)
is the Euclidean distance between \(\mathbf{x}_i\)
and the mean of its cluster \(\bar{\mathbf{c}}_g\)
.
对于single run 而言是 算出每组的 Euclidean distance 的平方然后根据组内有多少da ta
data相加 , 但是如何算2500组的SSR呢