如何使用sklearn.inspection.permutation_importance实现聚类算法

Question

import numpy as np
from sklearn.datasets import make_classification
from sklearn.cluster import KMeans

X, y = make_classification(n_samples=1000,
                           n_features=4,
                           n_informative=3,
                           n_redundant=0,
                           n_repeated=0,
                           n_classes=2,
                           random_state=0,
                           shuffle=False)

km = KMeans(n_clusters=3).fit(X)

result = permutation_importance(km, X, y, scoring='homogeneity_score', n_repeats=10, random_state=0, n_jobs=-1)
result

在真正的问题中，我没有y(真实标签)，我尝试使用y=None来使其成为一种无监督学习。但它不起作用。我得到了：

---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
<ipython-input-72-81045ae9cb66> in <module>()
----> 1 result = permutation_importance(km, X, y=None, scoring='homogeneity_score', n_repeats=10, random_state=0, n_jobs=-1)

5 frames
/usr/local/lib/python3.6/dist-packages/sklearn/metrics/cluster/_supervised.py in check_clusterings(labels_true, labels_pred)
     53     if labels_true.ndim != 1:
     54         raise ValueError(
---> 55             "labels_true must be 1D: shape is %r" % (labels_true.shape,))
     56     if labels_pred.ndim != 1:
     57         raise ValueError(

ValueError: labels_true must be 1D: shape is ()

有谁知道如何在没有真正标签的情况下实现？

Answer 1

首先，证明k-means对于特征的排列是不变的，这是很容易的。因为和是排列不变的。

如果你仍然想进行实验，可以尝试使用一个0数组作为y？