kmeans算法python实例，用于数据分类，以足球成绩为数据

               

name = ['china','japan','korea','iran','saudi','lraq','qatar','arab',         'uzbekistan','thailand','vietnam','oman','bahrain','northkorea','indonesia'] # 将国家名字保存在列表，以便调用 country =[[50, 50, 9], [28, 9, 4], [17, 15, 3], [25, 40, 5], [28, 40, 2], [50, 50, 1],           [50, 40, 9], [50, 40, 9], [40, 40, 5], [50, 50, 9], [50, 50, 9], [50, 50, 9],           [40, 40, 9], [40, 32, 17], [50, 50, 9]] # 将国家成绩保存为列表，以便调用 word1 = [] word2 = [] asia = [] groupi = [] groupj = [] groupk = [] namei = [] namej = [] namek = [] for i in range(15):     word1.append(country[i][0])     word2.append(country[i][1])     asia.append(country[i][2]) # 把2006,2010,2007年的排名读入列表word1，word2，asia中 for i in range(15):     word1[i] = (word1[i] - min(word1)) / (max(word1) - min(word1)) for i in range(15):     word2[i] = (word2[i] - min(word2)) / (max(word2) - min(word2)) for i in range(15):     asia[i] = (asia[i] - min(asia)) / (max(asia) - min(asia)) # 将每个年度的数据归一化 for i in range(15):     country[i][0] = word1[i]     country[i][1] = word2[i]     country[i][2] = asia[i] # 将归一化后的数据放回原来的列表内 def kmeans(i, j, k,number,country):  # i,j,k表示以哪几个点为种子点,number表示第几次分组，country是成绩列表     namei.clear()     namej.clear()     namek.clear()     # 将上次的分组清空     dis = []     for m in range(15):         dis.append(pow(             pow(country[m][0] - country[i][0], 2) + pow(country[m][1] - country[i][1], 2) + pow(country[m][2] - country[i][2],                                                                                           2), 0.5))         # 计算country[]到中心点country[i]的距离，将距离添加到dis的列表里面         dis.append(pow(             pow(country[m][0] - country[j][0], 2) + pow(country[m][1] - country[j][1], 2) + pow(country[m][2] - country[j][2],                                                                                           2), 0.5))         dis.append(pow(             pow(country[m][0] - country[k][0], 2) + pow(country[m][1] - country[k][1], 2) + pow(country[m][2] - country[k][2],                                                                                           2), 0.5))         index = dis.index(min(dis))  # 找出最小值的下标         if index == 0:             groupi.append(country[m])  # 将最小值的那个国家添加到分组中去             namei.append(name[m])         if index == 1:             groupj.append(country[m])             namej.append(name[m])         if index == 2:             groupk.append(country[m])             namek.append(name[m])         dis.clear()     number = str(number)     print("第" + number + "次分组为:")     print(namei)     print(namej)     print(namek) def newcenter(group):  # 重新计算中心点     center0 = 0     center1 = 0     center2 = 0     center = []     for i in range(len(group)):  # 求和         center0 = group[i][0] + center0         center1 = group[i][1] + center1         center2 = group[i][2] + center2     center.append(center0 / len(group))  # 将新的中心点的三个数据添加到列表center     center.append(center1 / len(group))     center.append(center2 / len(group))     return center number = 1 country.append(country[1])  # 用country的16,17,18位来存放中心点 country.append(country[9]) country.append(country[12]) kmeans(15, 16, 17,number,country) nai = namei naj = namej nak = namek # 将分组读入nai，naj，nak，用于判断是否收敛 count = 1 while count <= 10:  # 当运行10次,结束循环，认为收敛     number = number+1     country[15] = newcenter(groupi)     country[16] = newcenter(groupj)     country[17] = newcenter(groupk)     groupi.clear()     groupj.clear()     groupk.clear()     kmeans(15, 16, 17,number,country)     if namei == nai and namej == naj and namek == nak:         number = number + 1         country[15] = newcenter(groupi)         country[16] = newcenter(groupj)         country[17] = newcenter(groupk)         groupi.clear()         groupj.clear()         groupk.clear()         kmeans(15, 16, 17, number, country)         if namei == nai and namej == naj and namek == nak:             break         # 如果这次和上次的分组一样，则在进行一次分组，如果还是一样，跳出循环，认为收敛     count = count + 1 结果：