纯python代码DBSCAN算法分类鸢尾花数据

好的，以下是纯Python代码实现DBSCAN算法对鸢尾花数据进行聚类分类：

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt

# 加载数据集
def load_dataset():
    iris = pd.read_csv('iris.csv')
    X = iris.iloc[:, [0, 1, 2, 3]].values
    return X

# 计算距离矩阵
def dist(a, b):
    return np.sqrt(np.sum((a - b) ** 2))

# DBSCAN算法实现
def dbscan(X, eps, min_pts):
    m = X.shape[0]
    visited = np.zeros(m)
    clusters = np.zeros(m)
    C = 0
    for i in range(m):
        if not visited[i]:
            visited[i] = 1
            N = get_neighbors(X, i, eps)
            if len(N) < min_pts:
                clusters[i] = -1
            else:
                C += 1
                clusters[i] = C
                expand_cluster(X, visited, clusters, i, N, C, eps, min_pts)
    return clusters

# 获取邻居点
def get_neighbors(X, i, eps):
    return np.where(np.array([dist(X[i], x) for x in X]) <= eps)[0]

# 扩展簇
def expand_cluster(X, visited, clusters, i, N, C, eps, min_pts):
    for j in N:
        if not visited[j]:
            visited[j] = 1
            N2 = get_neighbors(X, j, eps)
            if len(N2) >= min_pts:
                N = np.union1d(N, N2)
        if not clusters[j]:
            clusters[j] = C

# 可视化结果
def plot_result(X, clusters):
    colors = ['r', 'g', 'b', 'y', 'c', 'm']
    for i in range(len(clusters)):
        if clusters[i] == -1:
            plt.scatter(X[i][0], X[i][1], color='k', marker='.')
        else:
            plt.scatter(X[i][0], X[i][1], color=colors[int(clusters[i]) % len(colors)], marker='.')
    plt.show()

# 主函数
if __name__ == '__main__':
    X = load_dataset()
    eps = 0.5
    min_pts = 5
    clusters = dbscan(X, eps, min_pts)
    plot_result(X, clusters)

其中，`load_dataset()`函数用于加载鸢尾花数据集，`dist()`函数用于计算两个样本之间的距离，`get_neighbors()`函数用于获取邻居点，`expand_cluster()`函数用于扩展簇，`dbscan()`函数是DBSCAN算法的主要实现，`plot_result()`函数用于可视化聚类结果。

在主函数中，我们首先加载鸢尾花数据集，然后定义`eps`和`min_pts`两个参数，这两个参数分别表示半径和最小点数，用于确定核心对象和边界点。最后，我们调用`dbscan()`函数进行聚类，并利用`plot_result()`函数将聚类结果可视化出来。