4.K均值算法--应用

from sklearn.datasets import load_sample_image
from sklearn.cluster import KMeans
import matplotlib.pyplot as plt
import sys
import matplotlib.image as img
import numpy as np

# 读取一张图片
image = img.imread("./wh1.jpg")
print(‘原图片文件大小：‘, image.size)
print(‘原图片占内存大小：‘, sys.getsizeof(image))
print(‘原图片的数据结构：\n‘, image)

plt.rcParams[‘font.sans-serif‘] = [‘SimHei‘]
plt.title("原图片")
plt.imshow(image)
plt.show()

# 用kmeans对图片像素颜色进行聚类
image = image[::3, ::3]  
X = image.reshape(-1, 3)
print(image.shape, X.shape)
n_colors = 64
model = KMeans(n_colors)
labels = model.fit_predict(X)
colors = model.cluster_centers_  

new_image = colors[labels].reshape(image.shape)  
print(‘压缩图片文件大小：‘, new_image.size)
print(‘压缩图片占内存大小：‘, sys.getsizeof(new_image))
print(‘压缩图片的数据结构：\n‘, new_image)
img.imsave(‘D://机器学习/2.jpg‘, new_image)

plt.title("压缩图片")
plt.imshow(new_image.astype(np.uint8))  
plt.show()

plt.title("二次压缩图片")
plt.imshow(new_image.astype(np.uint8)[::3, ::3])  
plt.show()

import numpy as np
import matplotlib.pyplot as plt
from sklearn.cluster import KMeans
import pandas as pd

che = pd.read_csv(‘201706120008王华 已处理.csv‘, index_col=0) 
data1 = che[che[‘车辆型号‘] == ‘奔驰S级‘] 
test_data = data1[[‘上牌年份‘, ‘表显里程‘, ‘价格‘]]  
data = np.array(test_data)  

model = KMeans(n_clusters=3)  
model.fit(data)
y = model.predict(data)  
model.cluster_centers_

a = []
b = []
c = []
d = []

for i in range(len(data)):
    if y[i] == 0:
        a.append(data[i, :])
        a1 = np.array(a)
    elif y[i] == 1:
        b.append(data[i, :])
        b1 = np.array(b)
    else:
        c.append(data[i, :])
        c1 = np.array(c)

plt.rcParams[‘font.sans-serif‘] = [‘SimHei‘]
plt.scatter(a1[:, 1], a1[:, 0], s=55, c=‘#ed1941‘, marker="*")  
plt.scatter(b1[:, 1], b1[:, 0], s=50, c=‘#84bf96‘, marker="o")  
plt.scatter(c1[:, 1], c1[:, 0], s=50, c=‘#e0861a‘, marker="v") 

plt.legend([‘中价车‘, ‘高价车‘, ‘低价车‘])
plt.ylabel("上牌年份")
plt.xlabel("表显里程")
plt.title("二手奔驰S级价格与上牌年份和表显里程的关系")
plt.show()