调试类型错误：不可散列的类型：'numpy.ndarray'

Question

我正在研究 kmeans 聚类。我在网络上的一些可用参考资料的帮助下写下了一段代码，但是当我运行这段代码时，它会引发一个错误：

    Traceback (most recent call last):
  File "clustering.py", line 16, in <module>
    ds = df[np.where(labels==i)]
  File "/usr/lib/python2.7/dist-packages/pandas/core/frame.py", line 1678, in __getitem__
    return self._getitem_column(key)
  File "/usr/lib/python2.7/dist-packages/pandas/core/frame.py", line 1685, in _getitem_column
    return self._get_item_cache(key)
  File "/usr/lib/python2.7/dist-packages/pandas/core/generic.py", line 1050, in _get_item_cache
    res = cache.get(item)
TypeError: unhashable type: 'numpy.ndarray'

虽然，许多以前的线程都存在相同的错误，但在我的程序中没有可以处理此错误的单一解决方案。如何调试此错误？

我使用的代码：

from sklearn import cluster
import pandas as pd

df = [
[0.57,-0.845,-0.8277,-0.1585,-1.616],
[0.47,-0.14,-0.5277,-0.158,-1.716],
[0.17,-0.845,-0.5277,-0.158,-1.616],
[0.27,-0.14,-0.8277,-0.158,-1.716]]

df = pd.DataFrame(df,columns= ["a","b","c","d", "e"])

# df = pd.read_csv("cleaned_remove_cor.csv")

k = 3
kmeans = cluster.KMeans(n_clusters=k)
kmeans.fit(df)
labels = kmeans.labels_
centroids = kmeans.cluster_centers_
from matplotlib import pyplot
import numpy as np

for i in range(k):
    # select only data observations with cluster label == i
    ds = df[np.where(labels==i)]
    # plot the data observations
    pyplot.plot(ds[:,0],ds[:,1],'o')
    # plot the centroids
    lines = pyplot.plot(centroids[i,0],centroids[i,1],'kx')
    # make the centroid x's bigger
    pyplot.setp(lines,ms=15.0)
    pyplot.setp(lines,mew=2.0)
pyplot.show()

我的 DataFrame 的形状是 (8127x600)

Answer 1

我试过了，这对我有用，将 pandas df 转换为 numpy 矩阵：

df = df.as_matrix(columns= ["a","b","c","d", "e"])