DIANA鸢尾花聚类可视化python代码

以下是一个使用DIANA聚类算法对鸢尾花数据集进行聚类，并可视化聚类结果的Python代码示例：

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
from scipy.spatial.distance import pdist, squareform
from sklearn.datasets import load_iris

def diana(X, k):
    # 计算初始距离矩阵
    D = squareform(pdist(X))
    n = len(X)

    # 开始聚类
    clusters = [[i] for i in range(n)]
    while len(clusters) < k:
        max_d = 0
        to_merge = None
        # 找到距离最远的簇对
        for i in range(len(clusters)):
            for j in range(i+1, len(clusters)):
                d = 0
                for m in clusters[i]:
                    for n in clusters[j]:
                        d += D[m][n]
                d /= len(clusters[i]) * len(clusters[j])
                if d > max_d:
                    max_d = d
                    to_merge = (i, j)
        # 合并簇
        i, j = to_merge
        clusters[i] = clusters[i] + clusters[j]
        del clusters[j]

    # 返回簇的索引
    labels = np.zeros(n)
    for i, c in enumerate(clusters):
        for j in c:
            labels[j] = i

    return labels


# 加载鸢尾花数据集
iris = load_iris()
X = iris.data

# 使用DIANA聚类
labels = diana(X, 3)

# 可视化聚类结果
df = pd.DataFrame(X, columns=['sepal_length', 'sepal_width', 'petal_length', 'petal_width'])
df['label'] = labels
colors = plt.cm.Spectral(np.linspace(0, 1, len(df['label'].unique())))
for i, color in zip(df['label'].unique(), colors):
    plt.scatter(df[df['label'] == i]['sepal_length'], df[df['label'] == i]['sepal_width'], color=color)
plt.xlabel('sepal length')
plt.ylabel('sepal width')
plt.show()

这个代码将加载鸢尾花数据集，使用DIANA聚类算法将其分为3个簇，并将聚类结果可视化。可以看到，不同的簇被不同的颜色表示出来。其中，横轴表示花萼长度，纵轴表示花萼宽度。