我在scikit learn（python）中尝试了2个集群算法：亲和传播和像这里的DBSCAN：
http://scikit-learn.org/stable/auto_examples/cluster/plot_dbscan.html#example-cluster-plot-dbscan-py
http://scikit-learn.org/0.12/auto_examples/cluster/plot_affinity_propagation.html
唯一的区别是X，不是：
centers = [[1, 1], [-1, -1], [1, -1]]
X, labels_true = make_blobs(n_samples=300, centers=centers, cluster_std=0.5,
random_state=0)
但我的2张图片的特征（坚固性和第一次幽默）每75张图片：
>>>>X.shape
(75,2)
我的功能：
[[ 5.54680144e-01 3.79948392e-01]
[ 8.70443729e-01 1.77502207e-01]
[ 4.37622956e-01 2.83559236e-05]
[ 6.72172924e-01 3.00142568e-04]
[ 5.18932433e-01 1.39958013e-03]
[ 7.23982700e-01 5.24452603e-04]
[ 4.95001469e-01 2.95420975e-03]
[ 1.64133952e-01 7.88177340e-04]
[ 2.52497558e-01 4.91686002e-01]
[ 6.86538731e-01 4.55305317e-01]
[ 7.22193099e-01 4.80662983e-01]
[ 4.67890677e-01 7.09454979e-03]
[ 6.18924155e-01 1.07420039e-04]
[ 3.53696287e-01 3.93981592e-03]
[ 6.07385501e-01 1.06825487e-02]
[ 2.84123395e-01 6.52089407e-01]
[ 2.36429649e-01 3.27600328e-03]
[ 2.30763588e-01 4.83091787e-03]
[ 1.59765027e-01 2.78884462e-02]
[ 1.86748975e-01 9.09235560e-05]
[ 5.34793573e-01 3.76842998e-04]
[ 5.05045881e-01 4.88897253e-01]
[ 2.10951780e-01 3.02640539e-04]
[ 1.23797482e+00 1.32727245e-01]
[ 5.58317299e-01 4.41987578e-01]
[ 4.35031459e-01 3.83944154e-03]
[ 9.40625702e-01 9.31836183e-05]
[ 8.40339071e-01 1.33191263e-04]
[ 2.01581656e-01 4.10607399e-01]
[ 2.70476981e-01 1.40600316e-01]
[ 3.99294959e-01 1.62107396e-03]
[ 3.46751951e-01 2.02122284e-03]
[ 1.66176385e-01 1.71828687e-04]
[ 3.28497515e-01 7.21117062e-02]
[ 3.40640083e-01 5.52515091e-01]
[ 8.38141960e-01 2.64894985e-01]
[ 4.68464960e-01 3.36463731e-01]
...
我有两个相反的回应：
找到了75个用于亲和力的聚类
为dbscan找到了1个群集。
也许我要重新调整数值或者还有其他事可做？