使用K-means，层次聚类和DBSCAN对鸢尾花的聚类

 

import numpy as np
from sklearn.cluster import KMeans, AgglomerativeClustering, DBSCAN
from sklearn import datasets, metrics
from collections import Counter
import matplotlib.pyplot as plt
import time
np.random.seed()


def prepare_data():
    iris = datasets.load_iris()
    x = iris.data
    y = iris.target
    return x, y


def eva_kmeans(x, y):
    k_begin = time.time()
    kmeans = KMeans(init="random", n_clusters=3, random_state=0).fit(x, y)
    k_end = time.time() - k_begin
    pre_label = kmeans.labels_
    key = list(set(pre_label))  # 获取簇标签
    # 重新定义簇标签
    key_c = []
    for i in key:
        idex = np.where(pre_label == i)
        key_c.append(Counter(y[idex]).most_common(1)[0][0])
    for i in range(pre_label.shape[0]):
        idex = key.index(pre_label[i])
        pre_label[i] = key_c[idex]
    kmean_ei = metrics.accuracy_score(y, pre_label) * 100
    kmean_rt = k_end
    plt.scatter(x[:, 0], x[:, 1], c=kmeans.labels_)
    plt.show()
    return kmean_ei, kmean_rt


def eva_hierarchical(x, y):
    k_begin = time.time()
    hier = AgglomerativeClustering(n_clusters=3, affinity="euclidean", linkage="average").fit(x, y)
    k_end = time.time() - k_begin
    pre_label = hier.labels_
    key = list(set(pre_label))      # 获取簇标签
    # 重新定义簇标签
    key_c =[]
    for i in key:
        idex = np.where(pre_label == i)
        key_c.append(Counter(y[idex]).most_common(1)[0][0])
    for i in range(pre_label.shape[0]):
        idex = key.index(pre_label[i])
        pre_label[i] = key_c[idex]
    hier_ei = metrics.accuracy_score(y, pre_label)*100
    hier_rt = k_end
    plt.scatter(x[:, 0], x[:, 1], c=hier.labels_)
    plt.show()
    return hier_ei, hier_rt


def eva_dbscan(x, y):
    k_begin = time.time()
    dbs = DBSCAN(eps=0.46,min_samples=11).fit(x, y)
    k_end = time.time() - k_begin
    pre_label = dbs.labels_
    key = list(set(pre_label))  # 获取簇标签
    # 重新定义簇标签
    key_c = []
    for i in key:
        idex = np.where(pre_label == i)
        key_c.append(Counter(y[idex]).most_common(1)[0][0])
    for i in range(pre_label.shape[0]):
        idex = key.index(pre_label[i])
        pre_label[i] = key_c[idex]
    dbs_ei = metrics.accuracy_score(y, pre_label) * 100
    dbs_rt = k_end
    plt.scatter(x[:, 0], x[:, 1], c=dbs.labels_)
    plt.show()
    return dbs_ei, dbs_rt


if __name__ == "__main__":
    x, y = prepare_data()
    kmean_ei, kmean_rt = eva_kmeans(x, y)
    hier_ei, hier_rt = eva_hierarchical(x, y)
    dbs_ei, dbs_rt = eva_dbscan(x, y)
    print("method    ","#accuracy/%    ","run_time/s")
    print("K_means    ",kmean_ei,"    ",kmean_rt)
    print("hierarchical    ",hier_ei,"    ",hier_rt)
    print("dbscan    ",dbs_ei,"    ",dbs_rt)

准确率：

 

K_means

 

hierarchical

dbscan