I did clustering with k-means, but I haven't complete my project, now I have to evaluate the result of the k-means clustering, and I want to do that with the easiest way. does anyone have any suggestions and example?. Help me please. Thanks before.
Solved – the easiest way to evaluate k-means clustering
clusteringk-means
Best Answer
The easiest way is to print the sum of squares and compare with other tools.
Since k-means heuristically minimizes this value, it should be close to the ideal value. If the value is much higher, there is a bug.
There is no "true" solution in clustering. It's successful, if you discovered new knowledge, not just because you optimized some statistic.