K-means python

K-means测试

一、实验目的

加深聚类算法的理解
熟悉K-means

二、实验环境

Windows + Python3+
一台装有集成开发环境（IDE）—— PyCharm的计算机

三、实验内容

1.根据已有的数据来进行聚类。

四、代码填写

import random
import math
import numpy as np
import matplotlib.pyplot as plt
plt.ion()
plt.rcParams['figure.figsize'] = (10, 10)        # 图像显示大小
plt.rcParams['font.sans-serif']=['SimHei']   #防止中文标签乱码，还有通过导入字体文件的方法
plt.rcParams['axes.unicode_minus'] = False
plt.rcParams['lines.linewidth'] = 0.5   #设置曲线线条宽度
#创建KMeans类
class KMeans():
    def __init__(self):
        self.data=None
        self.k=2
        self.partition=[]
        self.dim=0
        self.centers=[]
#train函数，用于训练
    def train(self,classes=2,dim=2,trainData=None):
        self.data=trainData
        self.k=classes
        self.dim=dim
        random.shuffle(self.data)
        for i in range(self.k):
            self.centers.append([self.data[i]])
            self.partition.append([self.data[i]])
        while True:
            plt.clf()
            temp=self.centers
            for example in self.data:
#挑出最近的点
                best=self.Select(self.centers,example)
                self.partition[best].append(example)
            self.centers = self.Center(self.partition)
            print(temp)
            if temp==self.centers:
                print(self.centers)
                break
            px = [a[0] for a in self.partition[0]]
            py = [a[1] for a in self.partition[0]]
            kx = [a[0] for a in self.partition[1]]
            ky = [a[1] for a in self.partition[1]]
            cx = [a[0][0] for a in self.centers]
            cy = [a[0][1] for a in self.centers]
#画出迭代过程
            plt.scatter(px, py, c="red")
            plt.scatter(kx, ky, c="yellow")
            plt.scatter(cx, cy, c='blue')
            plt.pause(2)
            self.partition = [[], []]
#计算中心位置
    def Center(self,partition):
        centers=[]
        for points in partition:
            tt=[]
            temp=np.zeros(self.dim)
            for point in points:
                temp=temp+point
            center=temp/len(points)
            tt.append(list(center))
            centers.append(tt)
        return centers
#根据计算结果进行测试
    def Test(self,testData=None):
        result=[]
        for example in testData:
            best = self.Select(self.centers, example)
            result.append(best)
        print(result)         
#计算距离
    def calcuDis(self,a,b):
        sum=0
        for i in range(self.dim):
            sum=sum+pow(a[i]-b[i],2)
        return math.sqrt(sum)
#排序挑出最近的点
    def Select(self,data,example):
        centers=self.Center(data)
        disLis=[]
        for center in centers:
            disLis.append(self.calcuDis(center[0],example))
        ans = sorted(range(len(disLis)), key=lambda k: disLis[k])[0]
        return ans
#自定义生成点集
def generate(num):
    data=[]
    for i in range(num):
             x=list(np.random.randint(low=1,high=5*num)*np.random.randn(num,2))
        x=[list(i) for i in x]
        data.extend(x)
    return data

km=KMeans()
data=[[0.697,0.460],
      [0.774,0.376],
      [0.634,0.264],
      [0.608,0.318],
      [0.556,0.215],
      [0.403,0.237],
      [0.481,0.149],
      [0.437,0.211],
      [0.666,0.091],
      [0.243,0.267]]

km.train(classes=2,trainData=data,dim=2)
plt.ioff()  # 关闭画图的窗口，即关闭交互模式
plt.show()  # 显示图片，防止闪退

五、实验结果

给定数据测试

自定义数据集测试，动态演示

苏苏