Python实现K-Means聚类算法

一、算法简介

1、K-Means算法是硬聚类算法，是典型的基于原型的目标函数聚类方法的代表，它是数据点到原型的某种距离作为优化的目标函数，利用函数求极值的方法得到迭代运算的调整规则。

2、K-Means算法以欧式距离作为相似度测度，它是求对应某一初始聚类中心向量V最优分类，使得评价指标J最小。

3、算法采用误差平方和准则函数作为聚类准则函数。

 

二、算法过程

（1）从n个向量对象任意选择k个向量作为初始聚类中心

（2）根据在步骤（1）中设置的k个向量（中心对象向量），计算每个对象与这k个中心对象各自的距离。

（3）对于步骤（2）中的计算，任何一个向量与这k个向量都有一个距离，有的远有的近，把这个向量和距离它最近的中心向量对象归在一个类簇中。

（4）重新计算每个类簇的中心对象向量位置

（5）重复（3）、（4）两个步骤，直到类簇聚类方案中的向量归类变化极少为止。

 

三、环境准备

• Ubuntu 16.04
• Python 2.7
• 组件：NumPy、SciPy、Matplotlib、Scikit-learn(sklearn)

Ubuntu下Python组件的安装：

 

sudo apt-get install python-numpy

sudo apt-get install python-scipy

sudo apt-get install python-matplotlib

sudo apt-get install python-sklearn

 

四、数据准备

中国城市经纬度表（city.txt）

    【格式】省份与城市,纬度,经度

选取部分省份以及城市，此处选择了北京、上海、吉林、山东、安徽、云南、四川、宁夏回族自治区。（这些信息网络上都可以查到）

上海上海,121.48,31.22  
上海嘉定,121.24,31.4  
上海宝山,121.48,31.41  
上海川沙,121.7,31.19  
上海南汇,121.76,31.05  
上海奉贤,121.46,30.92  
上海松江,121.24,31  
上海金山,121.16,30.89  
上海青浦,121.1,31.15  
上海崇明,121.4,31.73  
云南昆明,102.73,25.04  
云南富民,102.48,25.21  
云南晋宁,102.58,24.68  
云南呈贡,102.79,24.9  
云南安宁,102.44,24.95  
云南昭通,103.7,29.32  
云南永善,103.63,28.22  
云南大关,103.91,27.74  
云南彝良,104.06,27.61  
云南鲁甸,103.54,27.21  
云南绥江,103.97,28.58  
云南盐津,104.28,28.08  
云南威信,105.05,27.85  
云南镇雄,104.86,27.42  
云南巧家,102.92,26.9  
云南永富,104.38,28.62  
云南曲靖,103.79,25.51  
云南宣威,104.09,26.24  
云南富源,104.24,25.67  
云南师宗,103.97,24.85  
云南嵩明,103.03,25.35  
云南会泽,103.27,26.41  
云南沽益,103.82,25.62  
云南罗平,104.3,24.88  
云南陆良,104.64,25.04  
云南宜良,103.12,24.9  
云南马龙,103.61,25.41  
云南路南,103.24,24.77  
云南寻甸,103.25,25.56  
云南玉溪,102.52,24.35  
云南华宁,102.93,24.26  
云南通海,102.75,24.09  
云南澄江,102.91,24.68  
云南江川,102.73,24.27  
云南易门,102.15,24.67  
云南元江,102,23.59  
云南新平,101.98,24.06  
云南峨山,102.38,24.16  
云南思茅,101,22.79  
云南普洱,101.03,23.07  
云南镇沅,100.88,23.9  
云南景东,100.82,24.42  
云南景谷,100.71,23.5  
云南黑江,101.71,23.4  
云南澜沦,99.97,22.55  
云南西盟,99.47,22.73  
云南江城,101.88,22.58  
云南孟连,99.55,22.32  
云南临沦,100.09,23.88  
云南云县,100.12,24.44  
云南镇康,99.02,23.92  
云南永德,99.25,24.03  
云南凤庆,99.92,24.58  
云南双江,99.85,23.45  
云南沧源,99.24,23.15  
云南耿马,99.41,23.56  
云南保由,99.18,25.12  
云南施甸,99.15,24.69  
云南腾冲,98.51,25.01  
云南昌宁,99.61,24.82  
云南龙陵,98.7,24.58  
云南丽江,100.25,26.86  
云南华坪,101.24,26.63  
云南永胜,100.76,26.71  
云南宁蒗,100.82,27.29  
云南文山,104.24,23.37  
云南广南,105.09,24.05  
云南西畴,104.68,23.42  
云南麻栗坡,104.71,23.12  
云南马关,104.4,23.01  
云南丘北,104.19,24.03  
云南砚山,104.35,23.62  
云南富宁,105.6,23.62  
云南个旧,102.43,23.35  
云南弥勒,103.43,24.41  
云南蒙自,103.41,23.36  
云南元阳,102.81,23.17  
云南红河,102.42,23.35  
云南石屏,102.48,23.73  
云南泸西,103.76,24.52  
云南金平,103.24,22.77  
云南开远,103.23,23.7  
云南绿春,102.42,23.01  
云南建水,102.79,23.64  
云南河口,103.98,22.52  
云南屏边,103.67,22.68  
云南景淇,100.79,22  
云南勐海,100.5,21.95  
云南勐腊,101.56,21.48  
云南楚雄,101.54,25.01  
云南元谋,101.85,25.7  
云南武定,102.36,25.55  
云南禄丰,102.08,25.15  
云南南华,101.26,25.21  
云南大姚,101.34,25.73  
云南永仁,101.7,26.07  
云南禄劝,102.45,25.58  
云南牟定,101.58,25.32  
云南双柏,101.67,24.68  
云南姚安,101.24,25.4  
云南下关,100.24,25.45  
云南剑川,99.88,26.53  
云南洱源,99.94,26.1  
云南宾川,100.55,25.82  
云南弥渡,100.52,25.34  
云南永平,99.52,25.45  
云南鹤庆,100.18,26.55  
云南大理,100.19,25.69  
云南漾濞,99.98,25.68  
云南云龙,99.39,25.9  
云南祥云,100.56,25.48  
云南巍山,100.33,25.23  
云南南涧,100.51,25.04  
云南潞西,98.6,24.41  
云南陇川,97.96,24.33  
云南盈江,97.93,24.69  
云南畹町,98.08,24.08  
云南瑞丽,97.83,24  
云南梁河,98.3,24.78  
云南泸水,98.82,25.97  
云南碧江,98.95,26.55  
云南福贡,98.92,26.89  
云南兰坪,99.29,26.49  
云南贡山,98.65,27.73  
云南中甸,99.72,27.78  
云南德钦,98.93,28.49  
云南维西,99.27,27.15  
北京北京,116.46,39.92  
北京平谷,117.1,40.13  
北京密云,116.85,40.37  
北京顺义,116.65,40.13  
北京通县,116.67,39.92  
北京怀柔,116.62,40.32  
北京大兴,116.33,39.73  
北京房山,115.98,39.72    
吉林长春,125.35,43.88  
吉林吉林,126.57,43.87  
吉林农安,125.15,44.45  
吉林德惠,125.68,44.52  
吉林榆树,126.55,44.83  
吉林九台,126.83,44.15  
吉林双阳,125.68,43.53  
吉林永吉,126.57,43.87  
吉林舒兰,126.97,44.4  
吉林蛟河,127.33,43.75  
吉林桦甸,126.72,42.97  
吉林磐石,126.03,42.93  
吉林延吉,129.52,42.93  
吉林汪清,129.75,43.32  
吉林珲春,130.35,42.85  
吉林图们,129.83,42.98  
吉林和龙,129,42.52  
吉林安图,128.3,42.58  
吉林敦化,128.18,43.35  
吉林通化,125.92,41.49  
吉林柳河,125.7,40.88  
吉林海龙,125.65,42.53  
吉林辉南,126.03,42.68  
吉林靖宇,126.8,42.38  
吉林浑江,126.4,41.97  
吉林抚松,127.27,42.33  
吉林集安,126.17,41.15  
吉林长白,128.17,41.43  
吉林四平,124.37,43.17  
吉林梨树,124.33,43.32  
吉林怀德,124.82,43.5  
吉林伊通,125.32,43.33  
吉林辽源,125.15,42.97  
吉林东丰,125.5,42.68  
吉林双辽,123.5,43.52  
吉林白城,122.82,45.63  
吉林大安,124.18,45.5  
吉林扶余,124.82,45.2  
吉林乾安,124.02,45  
吉林长岭,123.97,44.3  
吉林通榆,123.13,44.82  
吉林洮安,122.75,45.35   
四川成都,104.06,30.67  
四川金堂,104.32,30.88  
四川双流,104.94,30.57  
四川蒲江,103.29,30.2  
四川郫县,103.86,30.8  
四川新都,104.13,30.82  
四川来易,102.15,26.9  
四川盐边,101.56,26.9  
四川温江,103.81,30.97  
四川灌县,103.61,31.04  
四川彭县,103.94,30.99  
四川什邡,104.16,31.13  
四川广汉,104.25,30.99  
四川新津,103.78,30.42  
四川邛崃,103.47,30.42  
四川大邑,103.53,30.58  
四川崇庆,103.69,30.63  
四川绵阳,104.73,31.48  
四川江油,104.7,31.8  
四川青川,105.21,32.59  
四川平武,104.52,32.42  
四川光元,105.86,32.44  
四川旺苍,106.33,32.25  
四川剑阁,105.45,32.03  
四川梓潼,105.16,31.64  
四川三台,105.06,31.1  
四川盐亭,105.35,31.23  
四川射洪,105.31,30.9  
四川遂宁,105.58,30.52  
四川蓬溪,105.74,30.78  
四川中江,104.68,31.06  
四川德阳,104.37,31.13  
四川绵竹,104.19,31.32  
四川安县,104.41,31.64  
四川北川,104.44,31.89  
四川内江,105.04,29.59  
四川乐至,105.02,30.3  
四川安岳,105.3,30.12  
四川威远,104.7,29.57  
四川资中,104.85,29.81  
四川资阳,104.6,30.19  
四川简阳,104.53,30.38  
四川隆昌,105.25,29.64  
四川宜宾,104.56,29.77  
四川富顺,104.97,29.24  
四川南溪,104.96,28.87  
四川江安,105.06,28.71  
四川纳溪,105.38,28.77  
四川泸县,105.46,28.96  
四川合江,105.78,28.79  
四川泸州,105.39,28.91  
四川古蔺,105.79,28.03  
四川叙水,105.44,28.19  
四川长宁,104.91,28.6  
四川兴文,105.06,28.36  
四川琪县,104.81,28.38  
四川高县,104.52,28.4  
四川筠连,104.53,28.16  
四川屏由,104.15,28.68  
四川乐由,103.73,29.59  
四川夹江,103.59,29.75  
四川洪雅,103.38,29.95  
四川丹棱,103.53,30.04  
四川青神,103.81,29.86  
四川眉由,103.81,30.05  
四川彭由,103.83,30.22  
四川井研,104.06,29.67  
四川仁寿,104.09,30  
四川犍为,103.93,29.21  
四川沐川,103.98,28.96  
四川娥眉,103.5,29.62  
四川马边,103.53,28.87  
四川峨边,103.25,29.23  
四川金口,103.13,29.24  
四川涪陵,107.36,29.7  
四川垫江,107.34,30.36  
四川丰都,107.7,29.89  
四川石柱,108.13,29.98  
四川秀山,108.97,28.47  
四川西阳,108.75,28.85  
四川黔江,108.81,29.53  
四川彭水,108.19,29.29  
四川武隆,108.72,29.29  
四川南川,107.13,29.15  
四川万县,108.35,30.83  
四川开县,108.39,31.23  
四川城口,108.67,31.98  
四川巫溪,109.6,31.42  
四川巫山,109.86,31.1  
四川奉节,109.52,31.06  
四川云阳,108.89,30.99  
四川忠县,108.03,30.33  
四川梁平,107.78,30.66  
四川南允,106.06,30.8  
四川苍溪,105.96,31.75  
四川阆中,105.97,31.75  
四川仪陇,106.38,31.52  
四川南部,106.03,31.34  
四川西允,105.84,31.01  
四川营山,106.57,31.07  
四川蓬安,106.44,31.04  
四川广安,106.61,30.48  
四川岳池,106.43,30.55  
四川武胜,106.3,30.38  
四川华云,106.74,30.41  
四川达县,107.49,31.23  
四川万源,108.06,32.07  
四川宜汉,107.71,31.39  
四川开江,107.87,31.1  
四川邻水,106.91,30.36  
四川大竹,107.21,30.75  
四川渠县,106.94,30.85  
四川南江,106.83,32.36  
四川巴中,106.73,31.86  
四川平昌,107.11,31.59  
四川通江,108.24,31.95  
四川百沙,108.18,32  
四川雅安,102.97,29.97  
四川芦山,102.91,30.17  
四川名山,103.06,30.09  
四川荣经,102.81,29.79  
四川汉源,102.66,29.4  
四川石棉,102.38,29.21  
四川天全,102.78,30.09  
四川宝兴,102.84,30.36  
四川马尔康,102.22,31.92  
四川红原,102.55,31.79  
四川阿坝,101.72,31.93  
四川若尔盖,102.94,33.62  
四川黑水,102.95,32.06  
四川松潘,103.61,32.64  
四川南坪,104.19,33.23  
四川汶川,103.61,31.46  
四川理县,103.16,31.42  
四川小金,102.34,30.97  
四川金川,102.03,31.48  
四川壤塘,100.97,32.3  
四川茂汶,103.89,31.67  
四川康定,101.95,30.04  
四川炉霍,100.65,31.38  
四川甘孜,99.96,31.64  
四川新龙,100.28,30.96  
四川白玉,98.83,32.23  
四川德格,98.57,31.81  
四川石渠,98.06,33.01  
四川色达,100.35,32.3  
四川泸定,102.25,29.92  
四川丹巴,101.87,30.85  
四川九龙,101.53,29.01  
四川雅江,101,30.03  
四川道孚,101.14,30.99  
四川理塘,100.28,30.03  
四川乡城,99.78,28.93  
四川稻城,100.31,29.04  
四川巴塘,99,30  
四川得荣,99.25,28.71  
四川西昌,102.29,27.92  
四川昭觉,102.83,28.03  
四川甘洛,102.74,28.96  
四川雷波,103.62,28.21  
四川宁南,102.76,27.07  
四川会东,102.55,26.74  
四川会理,102.21,26.67  
四川德昌,102.15,27.4  
四川美姑,103.14,28.33  
四川金阳,103.22,27.73  
四川布拖,102.8,27.7  
四川普格,102.52,27.38  
四川喜德,102.42,28.33  
四川越西,102.49,28.66  
四川盐源,101.51,27.42  
四川冕宁,102.15,28.58  
四川木里,101.25,27.9    
天津天津,117.2,39.13  
天津宁河,117.83,39.33  
天津静海,116.92,38.93  
天津蓟县,117.4,40.05  
天津宝坻,117.3,39.75  
天津武清,117.05,39.4  
宁夏回族自治区银川,106.27,38.47  
宁夏回族自治区永宁,106.24,38.28  
宁夏回族自治区贺兰,106.35,38.55  
宁夏回族自治区石嘴山,106.39,39.04  
宁夏回族自治区平罗,106.54,38.91  
宁夏回族自治区陶乐,106.69,38.82  
宁夏回族自治区吴忠,106.21,37.99  
宁夏回族自治区同心,105.94,36.97  
宁夏回族自治区灵武,106.34,38.1  
宁夏回族自治区中宁,105.66,37.48  
宁夏回族自治区盐池,107.41,37.78  
宁夏回族自治区中卫,105.18,37.51  
宁夏回族自治区青铜峡,106.07,38.02  
宁夏回族自治区固原,106.28,36.01  
宁夏回族自治区西吉,105.7,35.97  
宁夏回族自治区泾源,106.33,35.5  
宁夏回族自治区海原,105.64,36.56  
宁夏回族自治区隆德,106.11,35.63  
安徽合肥,117.27,31.86  
安徽长丰,117.16,32.47  
安徽淮南,116.98,32.62  
安徽凤台,116.71,32.68  
安徽淮北,116.77,33.97  
安徽濉溪,116.76,33.92  
安徽芜湖,118.38,31.33  
安徽铜陵,117.82,30.93  
安徽蚌埠,117.34,32.93  
安徽马鞍山,118.48,31.56  
安徽安庆,117.03,30.52  
安徽宿州,116.97,33.63  
安徽宿县,116.97,33.63  
安徽砀山,116.34,34.42  
安徽萧县,116.93,34.19  
安徽吴壁,117.55,33.55  
安徽泗县,117.89,33.49  
安徽五河,117.87,33.14  
安徽固镇,117.32,33.33  
安徽怀远,117.19,32.95  
安徽滁州,118.31,32.33  
安徽嘉山,117.98,32.78  
安徽天长,119,32.68  
安徽来安,118.44,32.44  
安徽全椒,118.27,32.1  
安徽定远,117.68,32.52  
安徽凤阳,117.4,32.86  
安徽巢湖,117.87,31.62  
安徽巢县,117.87,31.62  
安徽肥东,117.47,31.89  
安徽含山,118.11,31.7  
安徽和县,118.37,31.7  
安徽无为,117.75,31.3  
安徽卢江,117.29,31.23  
安徽宣城,118.73,31.95  
安徽当涂,118.49,31.55  
安徽郎溪,119.17,31.14  
安徽广德,119.41,30.89  
安徽泾县,118.41,30.68  
安徽南陵,118.32,30.91  
安徽繁昌,118.21,31.07  
安徽宁国,118.95,30.62  
安徽青阳,117.84,30.64  
安徽屯溪,118.31,29.72  
安徽休宁,118.19,29.81  
安徽旌得,118.53,30.28  
安徽绩溪,118.57,30.07  
安徽歙县,118.44,29.88  
安徽祁门,117.7,29.86  
安徽黟县,117.92,29.93  
安徽太平,118.13,30.28  
安徽石台,117.48,30.19  
安徽桐城,116.94,31.04  
安徽纵阳,117.21,30.69  
安徽怀宁,116.63,30.41  
安徽望江,116.69,30.12  
安徽宿松,116.13,30.15  
安徽太湖,116.27,30.42  
安徽岳西,116.36,30.84  
安徽潜山,116.53,30.62  
安徽东至,116.99,30.08  
安徽贵池,117.48,30.66  
安徽六安,116.49,31.73  
安徽霍丘,116.27,32.32  
安徽寿县,116.78,32.57  
安徽肥西,117.15,31.7  
安徽舒城,116.94,31.45  
安徽霍山,116.32,31.38  
安徽金寨,115.87,31.67  
安徽阜阳,115.81,32.89  
安徽毫县,116.76,33.86  
安徽涡阳,116.21,33.49  
安徽蒙城,116.55,33.25  
安徽利辛,116.19,33.12  
安徽颖上,116.26,32.62  
安徽阜南,115.6,32.63  
安徽临泉,115.24,33.06  
安徽界首,115.34,33.24  
安徽太和,115.61,33.16  
山东济南,117,36.65  
山东历城,117.07,36.69  
山东长清,116.73,36.55  
山东章丘,117.53,36.72  
山东青岛,120.33,36.07  
山东崂山,120.42,36.15  
山东胶南,119.97,35.88  
山东即墨,120.45,36.38  
山东胶县,120,36.28  
山东淄博,118.05,36.78  
山东枣庄,117.57,34.86  
山东滕县,117.17,35.09  
山东东营,118.49,37.46  
山东垦利,118.54,37.59  
山东利津,118.25,37.49  
山东德州,116.29,37.45  
山东宁津,116.8,37.64  
山东乐陵,117.22,37.74  
山东商河,117.15,37.31  
山东济阳,117.2,36.97  
山东禹城,116.66,36.95  
山东夏津,116,36.95  
山东陵县,116.58,37.34  
山东庆云,117.37,37.37  
山东临邑,116.86,37.2  
山东齐河,116.76,36.79  
山东平原,116.44,37.16  
山东武城,116.08,37.2  
山东滨州,118.03,37.36  
山东滨县,117.97,37.47  
山东广饶,118.41,37.04  
山东桓台,118.12,36.95  
山东邹平,117.75,36.89  
山东阳信,117.58,37.65  
山东沾化,118.14,37.7  
山东博兴,118.12,37.12  
山东高青,117.66,37.18  
山东惠民,117.51,17.49  
山东无棣,117.58,37.73  
山东潍坊,119.1,36.62  
山东潍县,119.22,36.77  
山东平度,119.97,36.77  
山东诸城,119.42,35.99  
山东安丘,119.2,36.42  
山东临朐,118.53,36.5  
山东寿光,118.73,36.86  
山东昌邑,119.41,36.86  
山东高密,119.75,36.38  
山东五莲,119.2,35.74  
山东昌乐,118.83,36.69  
山东高都,118.47,36.69  
山东烟台,121.39,37.52  
山东牟平,121.59,37.38  
山东文登,122.05,37.2  
山东海阳,121.17,36.76  
山东莱阳,120.71,36.97  
山东栖霞,120.83,37.28  
山东掖县,119.93,37.18  
山东长岛,120.73,37.91  
山东威海,122.1,37.5  
山东福山,121.27,37.49  
山东荣成,122.41,37.16  
山东乳山,121.52,36.89  
山东莱西,120.53,36.86  
山东招远,120.38,37.35  
山东黄县,120.51,37.64  
山东蓬莱,120.75,37.8  
山东临沂,118.35,35.05  
山东沂水,118.64,35.78  
山东日照,119.46,35.42  
山东临沭,118.73,34.89  
山东仓山,118.03,34.84  
山东平邑,117.63,35.49  
山东沂源,118.17,36.18  
山东沂南,118.47,35.54  
山东营县,118.83,35.57  
山东莒南,118.83,35.17  
山东郯城,118.35,34.61  
山东费县,117.97,35.26  
山东蒙阴,117.95,35.7  
山东泰安,117.13,36.18  
山东莱芜,117.67,36.19  
山东肥城,116.76,36.24  
山东平阴,116.46,36.29  
山东新汶,117.67,35.86  
山东新泰,117.76,35.91  
山东宁阳,116.8,35.76  
山东东平,116.3,35.91  
山东济宁,116.59,35.38  
山东兖州,116.83,35.54  
山东泗水,117.27,35.65  
山东鱼台,116.65,35  
山东嘉祥,116.34,35.41  
山东汶上,116.49,35.71  
山东曲阜,116.98,35.59  
山东邹县,116.97,35.39  
山东微山,117.12,34.8  
山东金乡,116.32,35.07  
山东荷泽,115.43,35.24  
山东郓城,115.94,35.59  
山东巨野,116.08,35.38  
山东单县,116.07,34.82  
山东曹县,115.53,34.83  
山东鄄城,115.5,35.57  
山东梁山,116.1,35.8  
山东成武,115.88,34.97  
山东定陶,115.57,35.07  
山东东明,115.08,35.31  
山东聊城,115.97,36.45  
山东高唐,116.23,36.86  
山东东阿,116.23,36.32  
山东莘县,115.67,36.24  
山东临清,115.72,36.68  
山东茌平,116.27,36.58  
山东阳谷,115.78,36.11  
山东冠县,115.45,35.47  

 

五、代码编写

#coding=utf-8
import numpy as np
import matplotlib.pyplot as plt
from sklearn.cluster import KMeans

#从磁盘读取城市经纬度数据
X = []
f = open('city.txt')
for v in f:
   x.append([float(v.split(',')[1]),float(v.split(',')[2])])
#转换成numpy array
X = np.array(X)
#类簇的数量
n_clusters = 5
#现在把数据和对应的分类书放入聚类函数中进行聚类
cls = KMeans(n_clusters).fit(X)
#X中每项所属分类的一个列表
cls.labels_
#画图
markers = ['^','x','o','*','+']
for i in range(n_clusters):
  members = cls.labels_ == i
  plt.scatter(X[members,0],X[members,1],s=60,marker=markers[i],c='b',alpha=0.5)
plt.title(' ')
plt.show()

六、测试结果

 

参考文献：《白话大数据与机器学习》，高扬、卫峥、尹会生编著，机械工业出版社