Davies Bouldin index

96 阅读 0 评论 64 点赞

我是靠谱客的博主优秀咖啡豆，这篇文章主要介绍Davies Bouldin index，现在分享给大家，希望可以做个参考。

The Davies–Bouldin index (DBI) is a metric for evaluating clustering algorithms. ^[1]

Let

C i

be a cluster of vectors. Let

X j

be a vector on

C i

$s_i = sqrt[q]{ frac{1}{t_i} displaystylesum_{j=1}^{t_i}left|x_j-a_iright|^q}$

Here

A i

is the centroid of

C i

S i

is a measure of scatter within the cluster.

$m_{i,j} = sqrt[p]{displaystylesum_{k=1}^{n}left|a_{k,i}-a_{k,j}right|^p }$

M i, j

is a measure of separation between cluster

C i

and cluster

C j

a k, i

is the kth element of

A i

$r_{i,j} equiv frac{s_i+s_j}{m_{i,j}}$

$r_i equiv max_{j : i neq j} r_{i,j}$

$bar{r} equiv frac{1}{n}displaystylesum_{i=1}^n r_i$

[edit]Notes and references

^ Davies, D. L.; Bouldin, D. W. A cluster separation measure.IEEE Trans. Pattern Anal. Mach. Intelligence 1979, 1, 224–227. 论文见附件

转载于:https://blog.51cto.com/visionsky/291912

以上就是优秀咖啡豆最近收集整理的关于Davies Bouldin index的全部内容，更多相关Davies内容请搜索靠谱客的其他文章。

本图文内容来源于网友提供，作为学习参考使用，或来自网络收集整理，版权属于原作者所有。