SSE and Silhouette Value for Evaluate K-means

2 次查看（过去 30 天）
Shahrooz Andrea
Shahrooz Andrea 2017-12-16
回答： Ting Su 2017-12-20
Hi, I wrote a program that uses k-means for clustering data, my data is 2D and in range(-2, 2). I want to know how SSE and Silhouette Value works to find optimum "k"?? Besides I want to write a code and don't want to use built-in functions. Thanks

请先登录，再回答此问题。

回答（1 个）

Ting Su
Ting Su 2017-12-20
The silhouette value can be used to estimate the number of clusters. Function “evalclusters†also provides a few other criteria to estimate the optimal number of clusters. SSE is the measure that K-means tries to minimize when the number of clusters is given. It can’t be used to estimate the number of clusters because SSE values are monotonically non-decreasing as the numbers of clusters increases.

类别

在 Help Center 和 File Exchange 中查找有关 k-Means and k-Medoids Clustering 的更多信息

标签

产品

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by