k-means性能测试

Posted 2021-02-12 yjybupt

clf = MiniBatchKMeans(n_clusters=5000, batch_size=5000, n_init=1, max_iter=200, max_no_improvement=10).fit(names_vector)

主要测试参数：

n_init
max_iter
max_no_improvement

n_clusters=5000, batch_size=5000, n_init=1, max_iter=200, max_no_improvement=10

========Kmeans========
43.68166518211365
4275

 

n_clusters=5000, batch_size=5000, n_init=1, max_iter=100, max_no_improvement=10

========Kmeans========
40.18006610870361
4314

 

max_iter增加，时间会增加，但是增加的不明显

 

n_clusters=5000, batch_size=10000, n_init=1, max_iter=100, max_no_improvement=10