{"id":621,"date":"2019-06-22T15:54:35","date_gmt":"2019-06-22T07:54:35","guid":{"rendered":"http:\/\/101.32.245.93\/?p=621"},"modified":"2019-07-08T21:01:57","modified_gmt":"2019-07-08T13:01:57","slug":"%e8%81%9a%e7%b1%bb%e5%88%86%e6%9e%90%e7%9a%84%e8%bf%87%e7%a8%8b%e5%92%8c%e4%b8%a4%e4%b8%aa%e5%b8%b8%e7%94%a8%e7%9a%84%e8%81%9a%e7%b1%bb%e7%ae%97%e6%b3%95","status":"publish","type":"post","link":"http:\/\/www.ahomer.cn\/?p=621","title":{"rendered":"\u805a\u7c7b\u5206\u6790\u7684\u8fc7\u7a0b\u548c\u4e24\u4e2a\u5e38\u7528\u7684\u805a\u7c7b\u7b97\u6cd5"},"content":{"rendered":"

\u805a\u7c7b\u5206\u6790\u8fc7\u7a0b<\/h3>\n

\u4e00\u822c\u805a\u7c7b\u5206\u6790\u7684\u6570\u636e\u6e90\u662f\u9700\u8981\u76f8\u5bf9\u5e72\u51c0\u7684\uff0c\u5373\u9700\u8981\u505a\u7edf\u4e00\u7684\u7279\u5f81\u6e05\u6d17\u3001\u7279\u5f81\u53d8\u6362\u8fc7\u7a0b\uff0c\u5373\u7a7a\u503c\u3001\u975e\u6cd5\u503c\u3001\u5f02\u5e38\u503c\u3001\u7c7b\u522b\u53d8\u91cf\u7b49\u7684\u5904\u7406\u3002\u4e3b\u8981\u8fc7\u7a0b\u5982\u4e0b\uff1a
\n\"\"<\/p>\n

\u6570\u636e\u91c7\u96c6\uff1a<\/strong>\u6211\u4eec\u53ef\u4ee5\u8ba4\u4e3a\u662f\u7edf\u4e00\u7684ETL\u8fc7\u7a0b\uff0c\u8fd9\u91cc\u6d89\u53ca\u57cb\u70b9\u3001\u8f6c\u53d1\u3001\u5b58\u50a8\u3001\u63d0\u53d6\u7b49\u8fc7\u7a0b\u3002\u662f\u5178\u578b\u7684\u6570\u636e\u5206\u6790\u524d\u7f6e\u8fc7\u7a0b\u3002<\/p>\n

\u7279\u5f81\u53d8\u6362\/\u7279\u5f81\u9009\u62e9\uff1a<\/strong>\u805a\u7c7b\u5bf9\u5f02\u5e38\u6570\u636e\u7279\u522b\u654f\u611f\uff0c\u540c\u65f6\u539f\u59cb\u6570\u636e\u76f4\u63a5\u8fdb\u5165\u805a\u7c7b\u5206\u6790\u4e0d\u5927\u73b0\u5b9e\u3002\u7279\u5f81\u5904\u7406\u5305\u542b\u884c\u7ef4\u5ea6\u3001\u5217\u7ef4\u5ea6\u7684\u5904\u7406\uff0c\u884c\u7ef4\u5ea6\u4e3b\u8981\u5305\u62ec\uff1a\u7a7a\u503c\u3001\u975e\u6cd5\u503c\u3001\u5f02\u5e38\u503c\u7b49\u65b9\u9762\u7684\u5904\u7406\uff0c\u800c\u5217\u7ef4\u5ea6\u6d89\u53ca\u964d\u7ef4\u5904\u7406\uff0c\u5197\u4f59\u7684\u5217\u5bf9\u805a\u7c7b\u5f71\u54cd\u8f83\u5927\uff0c\u6240\u4ee5\u4e00\u822c\u805a\u7c7b\u5206\u6790\u4e4b\u524d\u4f1a\u505a\u4e00\u6b21PCA\u3002<\/p>\n

\u805a\u7c7b\u5206\u6790\u3001\u805a\u7c7b\u8bc4\u4f30\uff1a<\/strong>\u4e0b\u6587\u5c06\u91cd\u70b9\u4e3e\u4f8b\u63cf\u8ff0<\/p>\n

\u7ed3\u679c\u89e3\u8bfb\uff1a<\/strong>\u65e0\u76d1\u7763\u5b66\u4e60\uff0c\u4e00\u822c\u6ca1\u6709label\uff0c\u805a\u7c7b\u7684\u6570\u91cf\u4e5f\u662f\u672a\u77e5\u7684\uff0c\u9700\u8981\u7ed3\u5408\u4e1a\u52a1\u77e5\u8bc6\u8fdb\u4e00\u6b65\u89e3\u8bfb\u805a\u7c7b\u7ed3\u679c\uff0c\u6bd4\u5982\u4ece\u7528\u6237\u753b\u50cf\u7ef4\u5ea6\u8fdb\u4e00\u6b65\u5207\u5206\u540c\u4e2a\u5206\u7c7b\u7684\u6570\u636e\uff0c\u4ece\u7edf\u8ba1\u7ef4\u5ea6\u6316\u6398\u7279\u5f81\u3002<\/p>\n

\u77e5\u8bc6\u53d1\u73b0\uff1a<\/strong>\u89e3\u8bfb\u7ed3\u679c\u4e4b\u540e\uff0c\u9700\u8981\u843d\u5730\u5b9e\u73b0\u6216\u8005\u8f93\u51fa\u62a5\u544a\uff0c\u6211\u4eec\u628a\u8fd9\u4e2a\u8fc7\u7a0b\u79f0\u4e3a\u77e5\u8bc6\u53d1\u73b0\u7684\u8fc7\u7a0b\u3002\u805a\u7c7b\u7ed3\u679c\u4ea7\u751f\u7684label\u7ed3\u679c\uff0c\u5f80\u5f80\u53ef\u4ee5\u4f5c\u4e3a\u76d1\u7763\u5b66\u4e60\u7684\u6765\u6e90\u3002<\/p>\n

\u4e0b\u9762\u4e3e\u4e24\u4e2a\u7ecf\u5178\u7684\u805a\u7c7b\u5206\u6790\u7b97\u6cd5\u8fdb\u4e00\u6b65\u8bf4\u660e\u3002<\/p>\n

Kmeans<\/h3>\n

\u57fa\u4e8e\u8ddd\u79bb\u805a\u7c7b\u7684\u805a\u7c7b\u7b97\u6cd5<\/p>\n

\u7b97\u6cd5\u6b65\u9aa4\uff1a<\/strong>
\n1\u3001\u6839\u636e\u8bbe\u5b9a\u5206\u7c7b\u6570\u91cf\uff0c\u968f\u673a\u751f\u6210N\u4e2a\u4e2d\u5fc3\u70b9
\n2\u3001\u6bcf\u4e2a\u70b9\u8ba1\u7b97\u4e0e\u4e2d\u5fc3\u70b9\u8ddd\u79bb\uff0c\u6309\u6700\u8fd1\u8ddd\u79bb\u5408\u5e76\u5206\u7c7b
\n3\u3001\u57fa\u4e8e2\u91cd\u65b0\u8ba1\u7b97\u6bcf\u4e2a\u5206\u7c7b\u7684\u4e2d\u5fc3\u70b9
\n4\u3001\u91cd\u590d2~3\uff0c\u76f4\u5230\u4e2d\u5fc3\u70b9\u6536\u655b
\n<\/p>\n

sklearn\u5b9e\u73b0<\/strong><\/p>\n

from sklearn.cluster import KMeans\n\nkmeans = KMeans(n_clusters=3)\npred = kmeans.fit_predict(X)<\/code><\/pre>\n
<\/p>\n
\u805a\u7c7b\u6548\u679c<\/strong>
\n\"\"<\/p>\n
GMM<\/h3>\n
\u9ad8\u65af\u6df7\u5408\u6a21\u578b\u662f\u51e0\u4e2a\u9ad8\u65af\u5206\u5e03\u7684\u53e0\u52a0\uff0c\u6bcf\u4e00\u4e2a\u5206\u5e03\u4ee3\u8868\u4e00\u4e2a\u5206\u7c7b
\n<\/p>\n
\u7b97\u6cd5\u6b65\u9aa4\uff1a<\/strong>
\n1\u3001\u8bbe\u5b9a\u5206\u7c7b\u6570\u91cf
\n2\u3001\u5bf9\u6bcf\u4e00\u4e2a\u9ad8\u65af\u5206\u5e03\u5747\u503c\u3001\u65b9\u5dee\u968f\u673a\u521d\u59cb\u5316
\n3\u3001\u8ba1\u7b97\u6bcf\u4e2a\u6837\u672c\u7684\u5728\u5404\u4e2a\u9ad8\u65af\u5206\u5e03\u7684\u6982\u7387\u3001\u6743\u91cd\u503c\uff0cExpectation-step
\n4\u3001\u6839\u636e\u6700\u5927\u4f3c\u7136\u91cd\u65b0\u4f30\u7b97\u9ad8\u65af\u5206\u5e03\u5747\u503c\u3001\u65b9\u5dee\uff0cMaximization-step
\n5\u3001\u91cd\u590d3~4\u76f4\u5230\u9ad8\u65af\u5206\u5e03\u5747\u503c\u548c\u65b9\u5dee\u6536\u655b
\n<\/p>\n
sklearn\u5b9e\u73b0<\/strong><\/p>\n
from sklearn.mixture import GaussianMixture\n\ngmm = GaussianMixture(n_components=3)\ngmm = gmm.fit(X)\npred_gmm = gmm.predict(X)<\/code><\/pre>\n
<\/p>\n
\u805a\u7c7b\u6548\u679c<\/strong>
\n\"\"
\n<\/p>\n
\u6a21\u578b\u8bc4\u4ef7<\/h3>\n
GMM\u548cKMeans\u5206\u7c7b\u7ed3\u679c\u6709\u4e00\u5b9a\u7684\u968f\u673a\u6027\uff0c\u53ef\u80fd\u5f97\u5230\u5c40\u90e8\u4f18\u5316\u89e3\uff0c\u9700\u8981\u591a\u6b21\u5fae\u8c03\u3002
\nGMM\u662f\u6982\u7387\u5206\u7c7b\u7ed3\u679c\uff0c\u53ef\u4ee5\u5bf9\u6bd4topN\u5206\u7c7b\uff0c\u6bd4\u5982\u6587\u6863\u805a\u7c7b\uff1bGMM\u8ba1\u7b97\u4ee3\u4ef7\u66f4\u5927\uff0c\u5927\u6570\u636e\u96c6\u662f\u4e2a\u8d1f\u62c5\u3002
\n<\/p>\n
\u8bc4\u4f30\u6307\u6807<\/h3>\n
\u805a\u7c7b\u7b97\u6cd5\u7684\u8bc4\u4f30\u4e3b\u8981\u8861\u91cf\u4e24\u4e2a\u65b9\u9762\uff1a\u540c\u7c7b\u7d27\u5bc6\u7a0b\u5ea6\u3001\u7c7b\u95f4\u5206\u6563\u7a0b\u5ea6\uff1b\u4ece\u4e0d\u540c\u6307\u6807\u770b\uff0c\u540c\u7c7b\u8d8a\u7d27\u5bc6\u3001\u7c7b\u95f4\u8d8a\u5206\u6563\uff0c\u805a\u7c7b\u6548\u679c\u8d8a\u597d
\n\u57fa\u672c\u4e0a\u4e3a\u65e0\u76d1\u7763\u805a\u7c7b\uff0c\u6240\u4ee5\u201cRand Index\u201d\u4e4b\u7c7b\u7684\u6307\u6807\u4e0d\u662f\u592a\u5b9e\u7528\uff0c\u5e38\u7528EIBOW\u65b9\u6cd5 \uff0c\u6839\u636eSSE\u8bc4\u4f30\u805a\u7c7b\u6548\u679c\uff0c\u53e6\u5916\u5bf9\u4e8e\u805a\u7c7b\u7ed3\u679c\u8bc4\u4f30\u53ef\u4ee5\u91c7\u7528\u4ee5\u4e0b\u4e24\u79cd\u65b9\u5f0f\uff1a<\/p>\n
1\u3001Silhouette index<\/strong><\/p>\n
    s = (b-a)\/max(a,b)\n    ai\uff1a\u540c\u5206\u7c7b\u5185\uff0c\u5355\u70b9\u4e0e\u5176\u4ed6\u70b9\u7684\u5e73\u5747\u8ddd\u79bb\n    bi\uff1a\u5206\u7c7b\u4e2d\u7684\u70b9\u4e0e\u6700\u8fd1\u5206\u7c7b\u70b9\u7684\u5e73\u5747\u8ddd\u79bb\n    a\uff1a\u6240\u6709\u70b9ai\u7684\u5747\u503c\n    b\uff1a\u6240\u6709\u70b9bi\u7684\u5747\u503c<\/code><\/pre>\n
\u57fa\u4e8e\u8f6e\u5ed3\u7684\u8861\u91cf\u65b9\u6cd5\uff0c\u53ef\u80fd\u4e0d\u9002\u5408 Single link\u3001Complete link\u3001DBSCANE\u7b49\u805a\u7c7b\u65b9\u6cd5\u3002<\/p>\n
2\u3001Calinski-Harabasz<\/strong>
\nCH\u6307\u6807\u901a\u8fc7\u201c\u540c\u7c7b\u201d\u3001\u201c\u7c7b\u95f4\u201d\u534f\u65b9\u5dee\u77e9\u9635\u7684\u8ff9\uff08\u5404\u4e2a\u53d8\u91cf\u7684\u65b9\u5dee\uff09\u8861\u91cf\uff0c\u8ba1\u7b97\u901f\u5ea6\u8f83\u5feb\u3002<\/p>\n","protected":false},"excerpt":{"rendered":"
\u805a\u7c7b\u5206\u6790\u8fc7\u7a0b \u4e00\u822c\u805a\u7c7b\u5206\u6790\u7684\u6570\u636e\u6e90\u662f\u9700\u8981\u76f8\u5bf9\u5e72\u51c0\u7684\uff0c\u5373\u9700\u8981\u505a\u7edf\u4e00\u7684\u7279\u5f81\u6e05\u6d17\u3001\u7279\u5f81\u53d8\u6362\u8fc7\u7a0b\uff0c\u5373\u7a7a\u503c\u3001\u975e\u6cd5\u503c\u3001\u5f02\u5e38 […]<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1],"tags":[33,31,22],"class_list":["post-621","post","type-post","status-publish","format-standard","hentry","category-others","tag-33","tag-31","tag-22"],"_links":{"self":[{"href":"http:\/\/www.ahomer.cn\/index.php?rest_route=\/wp\/v2\/posts\/621","targetHints":{"allow":["GET"]}}],"collection":[{"href":"http:\/\/www.ahomer.cn\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"http:\/\/www.ahomer.cn\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"http:\/\/www.ahomer.cn\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"http:\/\/www.ahomer.cn\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=621"}],"version-history":[{"count":5,"href":"http:\/\/www.ahomer.cn\/index.php?rest_route=\/wp\/v2\/posts\/621\/revisions"}],"predecessor-version":[{"id":5360,"href":"http:\/\/www.ahomer.cn\/index.php?rest_route=\/wp\/v2\/posts\/621\/revisions\/5360"}],"wp:attachment":[{"href":"http:\/\/www.ahomer.cn\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=621"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"http:\/\/www.ahomer.cn\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=621"},{"taxonomy":"post_tag","embeddable":true,"href":"http:\/\/www.ahomer.cn\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=621"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}