如何绘制Kmeans输出的(10,6)维质心数组及解决索引报错问题

• 你的center是普通NumPy多维数组，仅支持整数、切片、布尔数组等类型的索引，不支持Pandas DataFrame专属的字符串列名索引语法center['cluster']，这是报错的核心原因。
• 额外注意你当前的clusterNumber是float类型，如果后续需要作为簇标签使用，建议先转换为整数类型，避免后续可视化等操作出现类型不兼容问题。

6维数据无法直接映射到2维平面，可选两种常用方案实现可视化： #

方案1：PCA降维后绘制2D散点图 #

通过PCA算法把6维特征压缩到2个主成分，再映射到平面绘制，代码如下：

import numpy as np
import matplotlib.pyplot as plt
from sklearn.decomposition import PCA

# 加载你提供的质心数组
centers = np.array([
    [19.6135, 19.8452, 19.9962, 20.1065, 20.1966, 20.2832],
    [26.5262, 29.6227, 31.4583, 32.7302, 33.7162, 34.6274],
    [13.3404, 13.268,  13.2414, 13.2246, 13.2134, 13.2087],
    [44.3025, 47.7419, 49.3674, 50.5635, 51.4984, 52.3669],
    [58.331,  63.568,  66.6059, 69.222,  71.03,   72.5983],
    [23.26,   25.2503, 26.5113, 27.3892, 28.0659, 28.6797],
    [38.6445, 42.4035, 44.3822, 45.5953, 46.591,  47.4789],
    [30.3485, 33.8124, 35.8269, 37.2325, 38.3075, 39.2721],
    [48.3545, 53.1971, 56.0548, 58.1482, 59.7034, 61.11],
    [34.8697, 38.4072, 40.2917, 41.5594, 42.5017, 43.3741]
])
# 将float类型的簇编号转换为整数
cluster_labels = clusterNumber.astype(int)

# PCA降维到2维
pca = PCA(n_components=2)
centers_2d = pca.fit_transform(centers)

# 绘制散点图
plt.figure(figsize=(8,6))
scatter = plt.scatter(centers_2d[:,0], centers_2d[:,1], c=cluster_labels, cmap='tab10', s=100)
# 添加簇编号标注
for i, label in enumerate(cluster_labels):
    plt.annotate(label, (centers_2d[i,0], centers_2d[i,1]), xytext=(5,5), textcoords='offset points')
plt.legend(handles=scatter.legend_elements()[0], labels=cluster_labels, title='簇编号')
plt.xlabel('PCA主成分1')
plt.ylabel('PCA主成分2')
plt.title('Kmeans质心降维可视化结果')
plt.show()

方案2：平行坐标图直接展示6维特征 #

不需要降维，可直接呈现每个簇在6个维度上的数值分布，代码如下：

import pandas as pd
import matplotlib.pyplot as plt
from pandas.plotting import parallel_coordinates

# 加载质心数组，转换为DataFrame并添加簇列
centers = np.array([
    [19.6135, 19.8452, 19.9962, 20.1065, 20.1966, 20.2832],
    [26.5262, 29.6227, 31.4583, 32.7302, 33.7162, 34.6274],
    [13.3404, 13.268,  13.2414, 13.2246, 13.2134, 13.2087],
    [44.3025, 47.7419, 49.3674, 50.5635, 51.4984, 52.3669],
    [58.331,  63.568,  66.6059, 69.222,  71.03,   72.5983],
    [23.26,   25.2503, 26.5113, 27.3892, 28.0659, 28.6797],
    [38.6445, 42.4035, 44.3822, 45.5953, 46.591,  47.4789],
    [30.3485, 33.8124, 35.8269, 37.2325, 38.3075, 39.2721],
    [48.3545, 53.1971, 56.0548, 58.1482, 59.7034, 61.11],
    [34.8697, 38.4072, 40.2917, 41.5594, 42.5017, 43.3741]
])
cluster_labels = clusterNumber.astype(int)
df = pd.DataFrame(centers, columns=[f'特征{i+1}' for i in range(6)])
df['cluster'] = cluster_labels

# 绘制平行坐标图
plt.figure(figsize=(10,6))
parallel_coordinates(df, 'cluster', colormap='tab10')
plt.title('质心平行坐标图（6维特征直接展示）')
plt.show()

内容的提问来源于stack exchange，提问作者david