宜配屋

本文实例讲述了Python聚类算法之基本K均值运算技巧。分享给大家供大家参考，具体如下：

基本K均值 ：选择 K 个初始质心，其中 K 是用户指定的参数，即所期望的簇的个数。每次循环中，每个点被指派到最近的质心，指派到同一个质心的点集构成一个。然后，根据指派到簇的点，更新每个簇的质心。重复指派和更新操作，直到质心不发生明显的变化。

# scoding=utf-8
import pylab as pl
points = [[int(eachpoint.split("#")[0]), int(eachpoint.split("#")[1])] for eachpoint in open("points","r")]
# 指定三个初始质心
currentCenter1 = [20,190]; currentCenter2 = [120,90]; currentCenter3 = [170,140]
pl.plot([currentCenter1[0]], [currentCenter1[1]],'ok')
pl.plot([currentCenter2[0]], [currentCenter2[1]],'ok')
pl.plot([currentCenter3[0]], [currentCenter3[1]],'ok')
# 记录每次迭代后每个簇的质心的更新轨迹
center1 = [currentCenter1]; center2 = [currentCenter2]; center3 = [currentCenter3]
# 三个簇
group1 = []; group2 = []; group3 = []
for runtime in range(50):
  group1 = []; group2 = []; group3 = []
  for eachpoint in points:
    # 计算每个点到三个质心的距离
    distance1 = pow(abs(eachpoint[0]-currentCenter1[0]),2) + pow(abs(eachpoint[1]-currentCenter1[1]),2)
    distance2 = pow(abs(eachpoint[0]-currentCenter2[0]),2) + pow(abs(eachpoint[1]-currentCenter2[1]),2)
    distance3 = pow(abs(eachpoint[0]-currentCenter3[0]),2) + pow(abs(eachpoint[1]-currentCenter3[1]),2)
    # 将该点指派到离它最近的质心所在的簇
    mindis = min(distance1,distance2,distance3)
    if(mindis == distance1):
      group1.append(eachpoint)
    elif(mindis == distance2):
      group2.append(eachpoint)
    else:
      group3.append(eachpoint)
  # 指派完所有的点后，更新每个簇的质心
  currentCenter1 = [sum([eachpoint[0] for eachpoint in group1])/len(group1),sum([eachpoint[1] for eachpoint in group1])/len(group1)]
  currentCenter2 = [sum([eachpoint[0] for eachpoint in group2])/len(group2),sum([eachpoint[1] for eachpoint in group2])/len(group2)]
  currentCenter3 = [sum([eachpoint[0] for eachpoint in group3])/len(group3),sum([eachpoint[1] for eachpoint in group3])/len(group3)]
  # 记录该次对质心的更新
  center1.append(currentCenter1)
  center2.append(currentCenter2)
  center3.append(currentCenter3)
# 打印所有的点，用颜色标识该点所属的簇
pl.plot([eachpoint[0] for eachpoint in group1], [eachpoint[1] for eachpoint in group1], 'or')
pl.plot([eachpoint[0] for eachpoint in group2], [eachpoint[1] for eachpoint in group2], 'oy')
pl.plot([eachpoint[0] for eachpoint in group3], [eachpoint[1] for eachpoint in group3], 'og')
# 打印每个簇的质心的更新轨迹
for center in [center1,center2,center3]:
  pl.plot([eachcenter[0] for eachcenter in center], [eachcenter[1] for eachcenter in center],'k')
pl.show()

运行效果截图如下：