MiniBatchKmeans 结果的实施问题
problem with implementation of MiniBatchKmeans result
在知道我使用 MiniBatchKmeans 后多次执行我的算法后是否可能有不同的簇大小？
也就是说：
cluster 1: size = 30 cluster 2: size = 24 cluster 3: size = 2
在重新执行小批量之后，
cluster 1: size = 15 cluster 2: size = 20 cluster 3: size = 21
kmeans = MiniBatchKMeans(n_clusters=nbK, init ='k-means++', max_iter=1000, max_no_improvement = 10)
kmeans.fit(X)
prediction = kmeans.predict(X)
您应该修复 random_state 以获得确定性结果。
kmeans = MiniBatchKMeans(n_clusters=nbK, init ='k-means++', max_iter=1000, max_no_improvement = 10, random_state=10)
查看文档：https://scikit-learn.org/stable/modules/generated/sklearn.cluster.MiniBatchKMeans.html
在知道我使用 MiniBatchKmeans 后多次执行我的算法后是否可能有不同的簇大小？ 也就是说：
cluster 1: size = 30 cluster 2: size = 24 cluster 3: size = 2
在重新执行小批量之后，
cluster 1: size = 15 cluster 2: size = 20 cluster 3: size = 21
kmeans = MiniBatchKMeans(n_clusters=nbK, init ='k-means++', max_iter=1000, max_no_improvement = 10)
kmeans.fit(X)
prediction = kmeans.predict(X)
您应该修复 random_state 以获得确定性结果。
kmeans = MiniBatchKMeans(n_clusters=nbK, init ='k-means++', max_iter=1000, max_no_improvement = 10, random_state=10)
查看文档：https://scikit-learn.org/stable/modules/generated/sklearn.cluster.MiniBatchKMeans.html