{"id":77,"date":"2025-09-01T18:33:10","date_gmt":"2025-09-01T10:33:10","guid":{"rendered":"http:\/\/www.zxlearning.space\/?p=77"},"modified":"2025-09-01T18:33:11","modified_gmt":"2025-09-01T10:33:11","slug":"%e6%9c%ba%e5%99%a8%e5%ad%a6%e4%b9%a0","status":"publish","type":"post","link":"http:\/\/www.zxlearning.space\/?p=77","title":{"rendered":"\u673a\u5668\u5b66\u4e60"},"content":{"rendered":"\n<div class=\"wp-block-jetpack-markdown\"><h2>\u4e00\u3001\u673a\u5668\u5b66\u4e60\u6982\u8ff0<\/h2>\n<h3>\uff08\u4e00\uff09\u673a\u5668\u5b66\u4e60\u57fa\u672c\u6982\u5ff5<\/h3>\n<ul>\n<li>\n<p>\u673a\u5668\u5b66\u4e60\u4e13\u95e8\u7814\u7a76\u5982\u4f55\u6a21\u62df\u6216\u5b9e\u73b0\u4eba\u7c7b\u7684\u5b66\u4e60\u884c\u4e3a\uff0c\u4f7f\u5176\u4e0d\u65ad\u5730\u6539\u5584\u81ea\u8eab\u7684\u6027\u80fd<\/p>\n<\/li>\n<li>\n<p>\u673a\u5668\u5b66\u4e60\u6a21\u578b=\u6570\u636e+\u673a\u5668\u5b66\u4e60\u7b97\u6cd5<\/p>\n<pre><code class=\"language-mermaid\">graph LR\nid0(&quot;\u65b0\u7684\u6570\u636e&quot;)--\u8f93\u5165--&gt;id1(&quot;\u6a21\u578b&quot;)--\u9884\u6d4b--&gt;id2(&quot;\u672a\u77e5\u5c5e\u6027&quot;)\nid3(&quot;\u5386\u53f2\u6570\u636e&quot;)--\u8bad\u7ec3--&gt;id1\n<\/code><\/pre>\n<\/li>\n<\/ul>\n<h3>\uff08\u4e8c\uff09\u673a\u5668\u5b66\u4e60\u7684\u65b9\u5f0f<\/h3>\n<h4>1.\u57fa\u4e8e\u89c4\u5219\u7684\u5b66\u4e60<\/h4>\n<ul>\n<li>\n<p>\u57fa\u4e8e\u89c4\u5219\u7684\u5b66\u4e60\u662f\u901a\u8fc7\u4eba\u5de5\u5bf9\u6570\u636e\u63d0\u53d6\u89c4\u5f8b\uff0c\u5c06\u89c4\u5f8b\u8f6c\u6362\u4e3a\u4ee3\u7801<\/p>\n<\/li>\n<li>\n<p>\u4f46\u90e8\u5206\u95ee\u9898\u65e0\u6cd5\u901a\u8fc7\u57fa\u4e8e\u89c4\u5219\u7684\u5b66\u4e60\u89e3\u51b3\uff1a<\/p>\n<ul>\n<li>\u56fe\u50cf\u548c\u8bed\u97f3\u8bc6\u522b<\/li>\n<li>\u81ea\u7136\u8bed\u8a00\u5904\u7406<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n<h4>2.\u57fa\u4e8e\u6a21\u578b\u7684\u5b66\u4e60<\/h4>\n<ul>\n<li>\u57fa\u4e8e\u6a21\u578b\u7684\u5b66\u4e60\u662f\u901a\u8fc7\u7f16\u5199\u673a\u5668\u5b66\u4e60\u7a0b\u5e8f\uff0c\u8ba9\u673a\u5668\u81ea\u5df1\u4ece\u6570\u636e\u4e2d\u63d0\u53d6\u89c4\u5f8b<\/li>\n<\/ul>\n<h3>\uff08\u4e09\uff09\u673a\u5668\u5b66\u4e60\u6570\u636e\u96c6<\/h3>\n<ul>\n<li>\n<p>\u6837\u672c\uff1a\u6570\u636e\u96c6\u4e2d\u7684\u4e00\u6761\u6570\u636e<\/p>\n<\/li>\n<li>\n<p>\u7279\u5f81\uff1a\u6837\u672c\u4e2d\u7684\u4e00\u4e2a\u5c5e\u6027<\/p>\n<\/li>\n<li>\n<p>\u76ee\u6807\u503c\uff1a\u6700\u7ec8\u8981\u9884\u6d4b\u7684\u5c5e\u6027<\/p>\n<\/li>\n<\/ul>\n<h2>\u4e8c\u3001\u673a\u5668\u5b66\u4e60\u7684\u5206\u7c7b<\/h2>\n<h3>\uff08\u4e00\uff09\u76d1\u7763\u5b66\u4e60<\/h3>\n<ul>\n<li>\u76d1\u7763\u5b66\u4e60\u662f\u6307\u4eba\u5de5\u4e3a\u673a\u5668\u63d0\u4f9b\u4e00\u5927\u5806\u6807\u8bb0\u597d\u7684\u6570\u636e<\/li>\n<li>\u673a\u5668\u901a\u8fc7\u5b66\u4e60\u5f52\u7eb3\u7b97\u6cd5\u6216\u6a21\u578b<\/li>\n<li>\u901a\u8fc7\u7b97\u6cd5\u6216\u6a21\u578b\u9884\u6d4b\u6570\u636e<\/li>\n<li>\u5e38\u89c1\u6a21\u578b\uff1aLinear regression\u3001Logistic regression\u3001SVM\u3001Neural netword<\/li>\n<\/ul>\n<h4>1.\u5206\u7c7b\u95ee\u9898<\/h4>\n<p>\u5206\u7c7b\u95ee\u9898\u662f\u76d1\u7763\u5b66\u4e60\u4e2d\u4e00\u4e2a\u6838\u5fc3\u95ee\u9898\uff1a<\/p>\n<ul>\n<li>\u5f53\u8f93\u51fa\u53d8\u91cfY\u53d6\u6709\u9650\u4e2a\u79bb\u6563\u503c\u65f6\uff0c\u9884\u6d4b\u95ee\u9898\u4fbf\u5f62\u6210\u4e86\u5206\u7c7b\u95ee\u9898<\/li>\n<li>\u76d1\u7763\u5b66\u4e60\u4ece\u6570\u636e\u4e2d\u5b66\u4e60\u4e00\u4e2a\u5206\u7c7b\u6a21\u578b\u6216\u5206\u7c7b\u51b3\u7b56\u51fd\u6570\uff0c\u5f62\u6210\u4e86\u5206\u7c7b\u5668<\/li>\n<li>\u5206\u7c7b\u5668\u5bf9\u4e8e\u8f93\u5165\u7684\u6570\u636e\uff0c\u6309\u7167\u5206\u7c7b\u6a21\u578b\u6216\u5206\u7c7b\u51b3\u7b56\u51fd\u6570\u8fdb\u884c\u5206\u7c7b<\/li>\n<\/ul>\n<h4>2.\u56de\u5f52\u95ee\u9898<\/h4>\n<ul>\n<li>\u56de\u5f52\u95ee\u9898\u662f\u6307\u5bf9\u4e00\u7cfb\u5217\u8fde\u7eed\u6027\u8f93\u51fa\u53d8\u91cf\u8fdb\u884c\u9884\u6d4b<\/li>\n<li>\u6570\u636e\u4f1a\u63d0\u4f9b\u5927\u91cf\u7684\u81ea\u53d8\u91cf\u548c\u8fde\u7eed\u56e0\u53d8\u91cf\uff0c\u901a\u8fc7\u5bfb\u627e\u81ea\u53d8\u91cf\u548c\u56e0\u53d8\u91cf\u4e4b\u95f4\u7684\u5173\u7cfb\uff0c\u4ece\u800c\u5f62\u6210\u9884\u6d4b\u6a21\u578b<\/li>\n<\/ul>\n<h3>\uff08\u4e8c\uff09\u65e0\u76d1\u7763\u5b66\u4e60<\/h3>\n<ul>\n<li>\u65e0\u76d1\u7763\u5b66\u4e60\u662f\u6307\u4eba\u5de5\u4e3a\u673a\u5668\u63d0\u4f9b\u7684\u6570\u636e\u4e2d\u6ca1\u6709\u5206\u7c7b\u6807\u8bb0\uff0c\u65e0\u76ee\u6807\u503c<\/li>\n<\/ul>\n<h4>1.\u805a\u7c7b\u95ee\u9898<\/h4>\n<ul>\n<li>\u805a\u7c7b\u95ee\u9898\u662f\u4e00\u79cd\u63a2\u7d22\u6027\u6570\u636e\u5206\u6790\u6280\u672f\uff0c\u5728\u6ca1\u6709\u4efb\u4f55\u76f8\u5173\u5148\u9a8c\u4fe1\u606f\u7684\u60c5\u51b5\u4e0b\uff0c\u5c06\u6570\u636e\u5212\u5206\u4e3a\u6709\u610f\u4e49\u7684\u5c0f\u7684\u7ec4\u522b\uff08\u4e5f\u79f0\u4e3a\u7c07\uff09<\/li>\n<li>\u7c07\u5185\u7684\u6210\u5458\u5177\u6709\u4e00\u5b9a\u7684\u76f8\u4f3c\u5ea6\uff0c\u7c07\u4e4b\u95f4\u5b58\u5728\u8f83\u5927\u5dee\u5f02<\/li>\n<\/ul>\n<h4>2.\u6570\u636e\u964d\u7ef4<\/h4>\n<ul>\n<li>\u65e0\u76d1\u7763\u964d\u7ef4\u662f\u6570\u636e\u7279\u5f81\u9884\u5904\u7406\u65f6\u4f7f\u7528\u7684\u6280\u672f\uff0c\u7528\u4e8e\u6e05\u9664\u6570\u636e\u4e2d\u7684\u566a\u58f0<\/li>\n<\/ul>\n<h3>\uff08\u4e09\uff09\u534a\u76d1\u7763\u5b66\u4e60<\/h3>\n<ul>\n<li>\u534a\u76d1\u7763\u5b66\u4e60\u662f\u6307\u6570\u636e\u4e2d\u65e2\u6709\u643a\u5e26\u6807\u8bb0\uff0c\u53c8\u6709\u4e0d\u643a\u5e26\u6807\u8bb0\u7684<\/li>\n<li>\u5728\u5904\u7406\u672a\u6807\u8bb0\u6570\u636e\u65f6\uff0c\u53ef\u4ee5\u91c7\u7528\u201c\u4e3b\u52a8\u5b66\u4e60\u201d\u65b9\u5f0f\uff1a\n<ul>\n<li>\u5229\u7528\u5df2\u6807\u8bb0\u6570\u636e\u8bad\u7ec3\u6a21\u578b<\/li>\n<li>\u5229\u7528\u8bad\u7ec3\u51fa\u7684\u6a21\u578b\u5957\u7528\u672a\u6807\u8bb0\u7684\u6a21\u578b<\/li>\n<li>\u901a\u8fc7\u8be2\u95ee\u9886\u57df\u4e13\u5bb6\u5206\u7c7b\u7ed3\u679c\u4e0e\u6a21\u578b\u5206\u7c7b\u7ed3\u679c\u8fdb\u884c\u5bf9\u6bd4\uff0c\u63d0\u9ad8\u548c\u6539\u5584\u6a21\u578b<\/li>\n<\/ul>\n<\/li>\n<li>\u4e5f\u53ef\u4ee5\u91c7\u7528\u201c\u805a\u7c7b\u5b66\u4e60\u201d\u65b9\u5f0f\n<ul>\n<li>\u901a\u8fc7\u805a\u7c7b\u5c06\u672a\u6807\u8bb0\u7684\u6570\u636e\u805a\u96c6\u5728\u4e00\u4e2a\u7c07\u4e2d\uff0c\u4ece\u800c\u4e3a\u672a\u6807\u8bb0\u7684\u6570\u636e\u6dfb\u52a0\u6807\u8bb0<\/li>\n<li>\u672c\u8d28\u4e0a\u5229\u7528\u4e86\u4e00\u4e2a\u5047\u8bbe\uff1a\u76f8\u4f3c\u7684\u6837\u672c\u62e5\u6709\u76f8\u4f3c\u7684\u8f93\u51fa<\/li>\n<\/ul>\n<\/li>\n<li>\u534a\u76d1\u7763\u5b66\u4e60\u53ef\u4ee5\u5206\u4e3a\u4e24\u7c7b\uff1a\n<ul>\n<li>\u7eaf\u534a\u76d1\u7763\u5b66\u4e60\uff1a\u672a\u6807\u8bb0\u6570\u636e\u4f5c\u4e3a\u8bad\u7ec3\u6570\u636e<\/li>\n<li>\u76f4\u63a8\u5b66\u4e60\uff1a\u672a\u6807\u4ef7\u6570\u636e\u4f5c\u4e3a\u9884\u6d4b\u6570\u636e<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n<h3>\uff08\u56db\uff09\u5f3a\u5316\u5b66\u4e60<\/h3>\n<h4>1.\u6982\u5ff5<\/h4>\n<ul>\n<li>\u5f3a\u5316\u5b66\u4e60\u662f\u673a\u5668\u5b66\u4e60\u7684\u5206\u652f\uff0c\u7528\u4e8e\u89e3\u51b3\u8fde\u7eed\u51b3\u7b56\u95ee\u9898<\/li>\n<li>\u5f3a\u5316\u5b66\u4e60\u7684\u76ee\u6807\u4e00\u822c\u662f\u53d8\u5316\u7684\u3001\u4e0d\u660e\u786e\u7684\uff0c\u751a\u81f3\u53ef\u80fd\u4e0d\u5b58\u5728\u7edd\u5bf9\u6b63\u786e\u7684\u6807\u7b7e<\/li>\n<\/ul>\n<h4>2.\u76ee\u6807<\/h4>\n<ul>\n<li>\n<p>\u5f3a\u5316\u5b66\u4e60\u7684\u76ee\u6807\u662f\u6784\u5efa\u4e00\u4e2a\u7cfb\u7edf\uff0c\u5728\u4e0e\u73af\u5883\u7684\u4ea4\u4e92\u4e2d\u63d0\u9ad8\u7cfb\u7edf\u7684\u6027\u80fd\uff0c\u73af\u5883\u4e00\u822c\u5305\u62ec\u53cd\u9988\u548c\u72b6\u6001<\/p>\n<\/li>\n<li>\n<p>\u7cfb\u7edf\u5728\u4e0e\u73af\u5883\u4ea4\u4e92\u7684\u8fc7\u7a0b\u4e2d\uff0c\u901a\u8fc7\u5f3a\u5316\u5b66\u4e60\u5f97\u5230\u4e00\u7cfb\u5217\u7684\u884c\u4e3a\uff0c\u901a\u8fc7\u63a2\u7d22\u6027\u7684\u8bd5\u9519\u6216\u8005\u501f\u52a9\u7cbe\u5fc3\u8bbe\u8ba1\u7684\u6fc0\u52b1\u7cfb\u7edf\u4f7f\u5f97\u6b63\u5411\u53cd\u9988\u6700\u5927\u5316<\/p>\n<pre><code class=\"language-mermaid\">graph LR\nid0(\u7cfb\u7edf)\nid1(\u73af\u5883)\nid1--\u53cd\u9988--&gt;id0\nid0--\u4ea4\u4e92--&gt;id1\nid1--\u72b6\u6001--&gt;id0\n<\/code><\/pre>\n<\/li>\n<\/ul>\n<h2>\u4e09\u3001\u62df\u5408\u95ee\u9898<\/h2>\n<h3>\uff08\u4e00\uff09\u6b20\u62df\u5408<\/h3>\n<h4>1.\u73b0\u8c61<\/h4>\n<ul>\n<li>\u6a21\u578b\u5728\u8bad\u7ec3\u96c6\u4e0a\u7684\u8868\u73b0\u6548\u679c\u5dee\uff0c\u6ca1\u6709\u5145\u5206\u5229\u7528\u6570\u636e<\/li>\n<li>\u9884\u6d4b\u7684\u51c6\u786e\u7387\u5f88\u4f4e\uff0c\u62df\u5408\u7ed3\u679c\u4e25\u91cd\u4e0d\u7b26\u5408\u9884\u671f<\/li>\n<\/ul>\n<h4>2.\u539f\u56e0<\/h4>\n<ul>\n<li>\u4ea7\u751f\u539f\u56e0\uff1a\u6a21\u578b\u8fc7\u4e8e\u7b80\u5355<\/li>\n<li>\u51fa\u73b0\u573a\u666f\uff1a\u4e00\u822c\u51fa\u73b0\u5728\u673a\u5668\u5b66\u4e60\u6a21\u578b\u65e9\u671f\u8bad\u7ec3\u9636\u6bb5<\/li>\n<\/ul>\n<h4>3.\u89e3\u51b3\u65b9\u6848<\/h4>\n<ul>\n<li>\u6dfb\u52a0\u5176\u4ed6\u7279\u5f81\u9879<\/li>\n<li>\u6dfb\u52a0\u591a\u9879\u5f0f\u7279\u5f81\uff1a\u5728\u7ebf\u6027\u6a21\u578b\u4e2d\u6dfb\u52a0\u9ad8\u6b21\u9879<\/li>\n<li>\u51cf\u5c11\u6b63\u5219\u5316\u53c2\u6570<\/li>\n<\/ul>\n<h3>\uff08\u4e8c\uff09\u8fc7\u62df\u5408<\/h3>\n<h4>1.\u73b0\u8c61<\/h4>\n<ul>\n<li>\u6a21\u578b\u5728\u8bad\u7ec3\u96c6\u4e0a\u8868\u73b0\u5f88\u597d\uff0c\u4f46\u5728\u6d4b\u8bd5\u96c6\u4e0a\u6548\u679c\u5f88\u5dee<\/li>\n<\/ul>\n<h4>2.\u539f\u56e0<\/h4>\n<ul>\n<li>\u4ea7\u751f\u539f\u56e0\uff1a\u6a21\u578b\u8fc7\u4e8e\u590d\u6742\u3001\u6570\u636e\u4e0d\u7eaf\u3001\u8bad\u7ec3\u6570\u636e\u592a\u5c11\u7b49<\/li>\n<li>\u51fa\u73b0\u573a\u666f\uff1a\u6a21\u578b\u4f18\u5316\u5230\u4e00\u5b9a\u7a0b\u5ea6<\/li>\n<\/ul>\n<h4>3.\u89e3\u51b3\u65b9\u6848<\/h4>\n<ul>\n<li>\u91cd\u65b0\u6e05\u6d17\u6570\u636e\uff1a\u51cf\u5c11\u6570\u636e\u4e0d\u7eaf<\/li>\n<li>\u589e\u5927\u8bad\u7ec3\u7684\u6570\u636e\u91cf<\/li>\n<li>\u91c7\u7528\u6b63\u5219\u5316\u5bf9\u53c2\u6570\u65bd\u52a0\u6743\u91cd<\/li>\n<li>\u91c7\u7528dropout\u65b9\u6cd5\u8fdb\u884c\u968f\u673a\u91c7\u6837\u8bad\u7ec3\u6a21\u578b<\/li>\n<\/ul>\n<h4>4.\u5965\u5361\u59c6\u5243\u5200\u539f\u5219<\/h4>\n<ul>\n<li>\u5965\u5361\u59c6\u5243\u5200\u539f\u5219\u6216\u79f0\u4e3a\u8282\u4fed\u539f\u5219\uff1a\u7ed9\u5b9a\u4e24\u4e2a\u5177\u6709\u76f8\u540c\u6cdb\u5316\u8bef\u5dee\u7684\u6a21\u578b\uff0c\u8f83\u7b80\u5355\u7684\u6a21\u578b\u6bd4\u590d\u6742\u7684\u6a21\u578b\u66f4\u53ef\u53d6<\/li>\n<li>\u6a21\u578b\u8d8a\u590d\u6742\uff0c\u51fa\u73b0\u8fc7\u62df\u5408\u7684\u6982\u7387\u8d8a\u9ad8<\/li>\n<\/ul>\n<h3>\uff08\u4e09\uff09\u6cdb\u5316\u80fd\u529b<\/h3>\n<ul>\n<li>\u8bad\u7ec3\u51fa\u7684\u6a21\u578b\u9002\u7528\u4e8e\u65b0\u6837\u672c\u7684\u80fd\u529b\uff0c\u79f0\u4e4b\u4e3a\u6cdb\u5316\u80fd\u529b<\/li>\n<li>\u6a21\u578b\u7684\u6cdb\u5316\u80fd\u529b\u5f3a\uff0c\u610f\u5473\u7740\u5bf9\u4e8e\u65b0\u6570\u636e\u6709\u5f88\u597d\u7684\u9002\u5e94\u80fd\u529b<\/li>\n<\/ul>\n<h2>\u56db\u3001KNN\u7b97\u6cd5<\/h2>\n<h3>\uff08\u4e00\uff09KNN\u7b97\u6cd5\u7b80\u4ecb<\/h3>\n<h4>1.\u6982\u5ff5<\/h4>\n<ul>\n<li>KNN\u7b97\u6cd5\u662f\u76d1\u7763\u5b66\u4e60\u5206\u7c7b\u7b97\u6cd5\uff0c\u4e3b\u8981\u89e3\u51b3\u73b0\u5b9e\u751f\u6d3b\u4e2d\u7684\u5206\u7c7b\u95ee\u9898<\/li>\n<li>KNN\u7b97\u6cd5\u7684\u601d\u60f3\u6781\u5ea6\u7b80\u5355\uff0c\u5e94\u7528\u6570\u5b66\u77e5\u8bc6\u5c11\uff0c\u4f46\u6548\u679c\u4e0d\u9519<\/li>\n<\/ul>\n<h4>2.\u8fc7\u7a0b<\/h4>\n<ul>\n<li>\u53d6\u4e00\u4e2ak\u503c\uff1a\u8868\u793a\u4e0e\u90bb\u8fd1\u7684k\u4e2a\u70b9\u8fdb\u884c\u6bd4\u8f83<\/li>\n<li>\u5728\u6240\u6709\u70b9\u4e2d\u627e\u5230\u79bb\u65b0\u6837\u672c\u6700\u8fd1\u7684k\u4e2a\u70b9<\/li>\n<li>\u88ab\u9009\u4e2d\u7684k\u4e2a\u70b9\u6839\u636e\u81ea\u8eab\u6240\u5c5e\u7c7b\u522b\u8fdb\u884c\u6295\u7968<\/li>\n<\/ul>\n<h4>3.\u539f\u7406<\/h4>\n<ul>\n<li>\n<p>\u8f93\u5165\uff1a\u8bad\u7ec3\u6570\u636e\u96c6\uff0cx\u4e3a\u5b9e\u4f8b\u7684\u7279\u5f81\u5411\u91cf\uff0cy\u4e3a\u5b9e\u4f8b\u6240\u5c5e\u7684\u7c7b\u522b\n$$\nT={(x_1,y_1),(x_2,y_2)\u2026(x_n,y_n)}\n$$<\/p>\n<\/li>\n<li>\n<p>\u8f93\u51fa\uff1a\u5b9e\u4f8bx\u6240\u5c5e\u7684\u7c7b\u522by<\/p>\n<\/li>\n<li>\n<p>\u6b65\u9aa4\uff1a<\/p>\n<ul>\n<li>\u9009\u62e9\u5408\u9002\u7684\u53c2\u6570k<\/li>\n<li>\u8ba1\u7b97\u672a\u77e5\u5b9e\u4f8b\u4e0e\u6240\u6709\u5df2\u77e5\u5b9e\u4f8b\u7684\u8ddd\u79bb<\/li>\n<li>\u9009\u62e9\u6700\u8fd1\u7684k\u4e2a\u5df2\u77e5\u5b9e\u4f8b<\/li>\n<li>\u6839\u636e\u5c11\u6570\u670d\u4ece\u591a\u6570\u7684\u539f\u5219\uff0c\u8ba9\u672a\u77e5\u5b9e\u4f8b\u5f52\u5c5e\u4e3ak\u4e2a\u5b9e\u4f8b\u4e2d\u6570\u91cf\u6700\u591a\u7684\u7c7b\u578b<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n<h3>\uff08\u4e8c\uff09\u8ddd\u79bb\u5ea6\u91cf\u7b97\u6cd5<\/h3>\n<p>\u5728\u673a\u5668\u5b66\u4e60\u7b97\u6cd5\u4e2d\uff0c\u7ecf\u5e38\u9700\u8981\u5224\u65ad\u4e24\u4e2a\u6837\u672c\u662f\u5426\u76f8\u4f3c\uff0c\u5e38\u7528\u7684\u65b9\u5f0f\u662f\u5c06\u76f8\u4f3c\u7684\u5224\u65ad\u8f6c\u6362\u4e3a\u8ddd\u79bb\u7684\u8ba1\u7b97<\/p>\n<h4>1.\u6b27\u5f0f\u8ddd\u79bb<\/h4>\n<p>\u6b27\u5f0f\u8ddd\u79bb\uff1a\u6307\u4e24\u70b9\u4e4b\u95f4\u901a\u8fc7\u76f4\u63a5\u8fde\u7ebf\u7684\u65b9\u5f0f\uff0c\u6240\u5f97\u5230\u7684\u7ebf\u6bb5\u7684\u957f\u5ea6\n$$\n\u4e8c\u7ef4\uff1a\u8ddd\u79bbL=\\sqrt{(x_1-x_2)^2+(y_1-y_2)^2}\\\\\n\u63a8\u5e7f\uff1a\u8ddd\u79bbd=\\sqrt{\\sum_{k=1}^{n}(x_{1k}-x_{2k})^2}\n$$<\/p>\n<h4>2.\u66fc\u54c8\u987f\u8ddd\u79bb<\/h4>\n<p>\u66fc\u54c8\u987f\u8ddd\u79bb\uff1a\u6307\u4e24\u70b9\u4e4b\u95f4\u6cbf\u7740\u5750\u6807\u8f74\u8fd0\u52a8\u7684\u8ddd\u79bb\uff08\u4f8b\u5982\u4e8c\u7ef4\u4e2d\uff0c\u6cbf\u7740x\u8f74\u79fb\u52a8\u7684\u8ddd\u79bb+\u6cbf\u7740y\u8f74\u79fb\u52a8\u7684\u8ddd\u79bb\uff09\n$$\n\u8ddd\u79bbd=\\sum_{k=1}^n\\ |x_{1k}-x_{2k}|\n$$<\/p>\n<h4>3.\u5207\u6bd4\u96ea\u592b\u8ddd\u79bb<\/h4>\n<p>\u5207\u6bd4\u96ea\u592b\u8ddd\u79bb\uff1a\u7c7b\u4f3c\u56fd\u9645\u8c61\u68cb\uff0c\u4e00\u4e2a\u68cb\u5b50\u53ef\u4ee5\u5411\u516b\u4e2a\u65b9\u5411\u79fb\u52a8\uff0c\u90a3\u4e48\u5b83\u4ece\u4e00\u4e2a\u68cb\u683c\u5230\u53e6\u4e00\u4e2a\u68cb\u683c\u7684\u6700\u5c0f\u8ddd\u79bb\n$$\n\u8ddd\u79bbd=max(|x_1-x_2|,|y_1-y_2|)\n$$<\/p>\n<h4>4.\u95f5\u5f0f\u8ddd\u79bb<\/h4>\n<p>\u95f5\u5f0f\u8ddd\u79bb\uff1a\u662f\u5bf9\u4e0a\u8ff0\u8ddd\u79bb\u7684\u6982\u62ec\u6027\u8868\u793a\n$$\n\u8ddd\u79bbd=\\sqrt[p]{\\sum_{k=1}^n |x_{1k}-x_{2k}|^p}\n$$<\/p>\n<ul>\n<li>\u5f53p=1\u65f6\uff0c\u8be5\u8ddd\u79bb\u4e3a\u66fc\u54c8\u987f\u8ddd\u79bb<\/li>\n<li>\u5f53p=2\u65f6\uff0c\u8be5\u8ddd\u79bb\u4e3a\u6b27\u5f0f\u8ddd\u79bb<\/li>\n<li>\u5f53p\u8d8b\u4e8e\u65e0\u7a77\u65f6\uff0c\u8be5\u8ddd\u79bb\u4e3a\u5207\u6bd4\u96ea\u592b\u8ddd\u79bb<\/li>\n<\/ul>\n<h3>\uff08\u4e09\uff09\u5f52\u4e00\u5316\u548c\u6807\u51c6\u5316<\/h3>\n<h4>1.\u9700\u6c42\u80cc\u666f<\/h4>\n<ul>\n<li>\u6837\u672c\u4e2d\u4f1a\u5177\u6709\u591a\u4e2a\u7279\u5f81\uff0c\u6bcf\u4e00\u4e2a\u7279\u5f81\u6709\u81ea\u5df1\u7684\u5b9a\u4e49\u57df\u548c\u53d6\u503c\u8303\u56f4\uff0c\u56e0\u6b64\u5b83\u4eec\u5bf9\u4e8e\u8ddd\u79bb\u8ba1\u7b97\u7684\u5f71\u54cd\u4e5f\u662f\u4e0d\u540c\u7684<\/li>\n<li>\u56e0\u6b64\u9700\u8981\u5bf9\u7279\u5f81\u8fdb\u884c\u5f52\u4e00\u5316\u5904\u7406\uff0c\u4f7f\u5f97\u7279\u5f81\u90fd\u7f29\u653e\u5230\u76f8\u540c\u533a\u95f4\u6216\u8005\u5206\u5e03\u5185<\/li>\n<\/ul>\n<h4>2.\u5f52\u4e00\u5316<\/h4>\n<ul>\n<li>\n<p>\u5f52\u4e00\u5316\uff1a\u901a\u8fc7\u5bf9\u539f\u59cb\u6570\u636e\u7684\u53d8\u6362\uff0c\u5c06\u6570\u636e\u6620\u5c04\u5230[0,1]\u533a\u95f4\u5185\n$$\nx^{\u2018}=\\frac{x-min}{max-min}\\\nx^{\u2019\u2018}=x^{\u2019}*(max-min)+min\n$$<\/p>\n<\/li>\n<li>\n<p>scikit-learn\u4e2d\u5b9e\u73b0\u5f52\u4e00\u5316\u7684API\uff1a<\/p>\n<pre><code class=\"language-python\">from sklearn.preprocessing import MinMaxScaler\n\n#1.\u51c6\u5907\u6570\u636e\ndata = [\u6570\u636e]\n\n#2.\u521d\u59cb\u5316\u5f52\u4e00\u5316\u5bf9\u8c61\ntransformer = MinMaxScaler()\n\n#3.\u8fdb\u884c\u5f52\u4e00\u5316\u5904\u7406\ndata = transformer.fit_transform(data)\n<\/code><\/pre>\n<\/li>\n<li>\n<p>\u6ce8\u610f\uff1a\u5f52\u4e00\u5316\u53d7\u5230\u6700\u5927\u503c\u548c\u6700\u5c0f\u503c\u7684\u5f71\u54cd\uff0c\u5bb9\u6613\u88ab\u5f02\u5e38\u6570\u636e\u5f71\u54cd\uff0c\u9002\u5408\u4f20\u7edf\u7cbe\u786e\u5c0f\u6570\u636e\u573a\u666f<\/p>\n<\/li>\n<\/ul>\n<h4>3.\u6807\u51c6\u5316<\/h4>\n<ul>\n<li>\n<p>\u6807\u51c6\u5316\n$$\nx^{&#8216;}=\\frac{x-mean}{\\sigma}\n$$<\/p>\n<ul>\n<li>\n<p>$ mean $\uff1a\u7279\u5f81\u7684\u5e73\u5747\u503c<\/p>\n<\/li>\n<li>\n<p>$ \\sigma $\uff1a\u7279\u5f81\u7684\u6807\u51c6\u5dee<\/p>\n<\/li>\n<\/ul>\n<\/li>\n<li>\n<p>scikit-learn\u4e2d\u5b9e\u73b0\u6807\u51c6\u5316\u7684API\uff1a<\/p>\n<pre><code class=\"language-python\">from sklearn.preprocessing import StandardScaler\n\n#1.\u51c6\u5907\u6570\u636e\ndata = [\u6570\u636e]\n\n#2.\u521d\u59cb\u5316\u6807\u51c6\u5316\u5bf9\u8c61\ntransformer = StandardScaler()\n\n#3.\u8fdb\u884c\u6807\u51c6\u5316\u5904\u7406\ndata = transformer.fit_transform(data)\n<\/code><\/pre>\n<\/li>\n<li>\n<p>\u6ce8\u610f\uff1a\u5bf9\u4e8e\u6807\u51c6\u5316\u6765\u8bf4\uff0c\u5982\u679c\u51fa\u73b0\u5f02\u5e38\u70b9\uff0c\u7531\u4e8e\u5177\u6709\u4e00\u5b9a\u7684\u6570\u636e\u91cf\uff0c\u5c11\u91cf\u5f02\u5e38\u70b9\u5bf9\u4e8e\u5e73\u5747\u503c\u7684\u5f71\u54cd\u4e0d\u5927<\/p>\n<\/li>\n<\/ul>\n<h3>\uff08\u56db\uff09KNN\u7b97\u6cd5API<\/h3>\n<h4>1.\u5bfc\u5305<\/h4>\n<pre><code class=\"language-python\">from sklearn.datasets import load_iris\nfrom sklearn.preprocessing import StandardScaler\nfrom sklearn.neighbors import KNeighborsClassifier\n<\/code><\/pre>\n<h4>2.\u52a0\u8f7d\u6570\u636e\u96c6<\/h4>\n<pre><code class=\"language-python\">#1.\u52a0\u8f7d\u5185\u7f6e\u7684\u9e22\u5c3e\u82b1\u6570\u636e\u96c6\niris = load_iris()\niris.data\n<\/code><\/pre>\n<h4>3.\u6570\u636e\u6807\u51c6\u5316<\/h4>\n<pre><code class=\"language-python\">#2.\u6570\u636e\u6807\u51c6\u5316\ntransformer = StandardScaler()\nx = transformer.fit_transform(iris.data)\n<\/code><\/pre>\n<h4>4.\u6a21\u578b\u8bad\u7ec3<\/h4>\n<pre><code class=\"language-python\">#3.\u6a21\u578b\u8bad\u7ec3\nestimator = KNeighborsClassifier(n_neighbors=3) #n_neighbors:K\u7684\u503c\nestimator.fit(x,iris.target)\n<\/code><\/pre>\n<h4>5.\u8fdb\u884c\u9884\u6d4b<\/h4>\n<pre><code class=\"language-python\">#4.\u5229\u7528\u6a21\u578b\u8fdb\u884c\u9884\u6d4b\nresult = estimator.predict(x)\nprint(result)\n<\/code><\/pre>\n<h3>\uff08\u4e94\uff09\u5206\u7c7b\u6a21\u578b\u8bc4\u4f30\u65b9\u6cd5<\/h3>\n<h4>1.\u6570\u636e\u5212\u5206<\/h4>\n<ul>\n<li>\u9700\u6c42\uff1a\u5982\u679c\u4f7f\u7528\u6240\u6709\u6570\u636e\u6765\u8bad\u7ec3\u6a21\u578b\uff0c\u5c06\u7f3a\u5c11\u7528\u4e8e\u6d4b\u8bd5\u7684\u6570\u636e\uff0c\u96be\u4ee5\u8bc4\u4f30\u6a21\u578b\u7684\u6cdb\u5316\u80fd\u529b<\/li>\n<li>\u6d4b\u8bd5\u96c6\uff1a\u4f7f\u7528\u6d4b\u8bd5\u96c6\u6765\u6d4b\u8bd5\u6a21\u578b\u5bf9\u4e8e\u65b0\u6837\u672c\u7684\u5224\u522b\u80fd\u529b\uff0c\u4ee5\u6d4b\u8bd5\u8bef\u5dee\u4f5c\u4e3a\u6cdb\u5316\u8bef\u5dee\u7684\u8fd1\u4f3c<\/li>\n<li>\u6d4b\u8bd5\u96c6\u7684\u8981\u6c42\uff1a\n<ul>\n<li>\u80fd\u591f\u4ee3\u8868\u6574\u4e2a\u6570\u636e\u96c6<\/li>\n<li>\u6d4b\u8bd5\u96c6\u4e0e\u8bad\u7ec3\u96c6\u4e92\u65a5<\/li>\n<li>\u6d4b\u8bd5\u96c6\u4e0e\u8bad\u7ec3\u96c6\u6bd4\u4f8b\u5efa\u8bae\uff1a2\u6bd48\u30013\u6bd47\u7b49<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n<h4>2.\u6570\u636e\u5212\u5206\u7684\u65b9\u6cd5<\/h4>\n<ul>\n<li>\u7559\u51fa\u6cd5\uff1a\u7b80\u5355\u5730\u5c06\u6570\u636e\u96c6\u5212\u5206\u4e3a\u8bad\u7ec3\u96c6\u548c\u6d4b\u8bd5\u96c6<\/li>\n<li>\u4ea4\u53c9\u9a8c\u8bc1\u6cd5\uff1a\u5c06\u6570\u636e\u96c6\u5212\u5206\u4e3a\u8bad\u7ec3\u96c6\u548c\u6d4b\u8bd5\u96c6\u4e4b\u540e\uff0c\u518d\u5c06\u8bad\u7ec3\u96c6\u5212\u5206\u4e3a\u8bad\u7ec3\u96c6\u548c\u9a8c\u8bc1\u96c6\uff08\u9a8c\u8bc1\u96c6\u7528\u4e8e\u8c03\u8bd5\u53c2\u6570\uff09<\/li>\n<li>\u7559\u4e00\u6cd5\uff1a\u6bcf\u6b21\u4ece\u8bad\u7ec3\u96c6\u4e2d\u62bd\u53d6\u4e00\u6761\u6570\u636e\u4f5c\u4e3a\u6d4b\u8bd5\u96c6<\/li>\n<li>\u81ea\u52a9\u6cd5\uff1a\u4ee5\u53ef\u653e\u56de\u91c7\u6837\u4e3a\u57fa\u7840\u968f\u673a\u62bd\u53d6m\u4e2a\u6837\u672c\uff0c\u6ca1\u88ab\u62bd\u5230\u7684\u5176\u4f59\u6570\u636e\u4f5c\u4e3a\u6d4b\u8bd5\u96c6<\/li>\n<\/ul>\n<h4>3.\u7559\u51fa\u6cd5<\/h4>\n<ul>\n<li>\n<p>\u5bfc\u5305<\/p>\n<pre><code class=\"language-python\">from sklearn.model_selection import train_test_split\nfrom sklearn.model_selection import StratifiedShuffleSplit\nfrom sklearn.model_selection import ShuffleSplit\nfrom collections import Counter\nfrom sklearn.datasets import load_iris\n<\/code><\/pre>\n<\/li>\n<li>\n<p>\u52a0\u8f7d\u6570\u636e\u96c6<\/p>\n<pre><code class=\"language-python\">#1.\u52a0\u8f7d\u6570\u636e\u96c6\nx,y = load_iris(return_X_y=True) \n#\u82e5\u4e0d\u52a0\u53c2\u6570\uff0c\u9ed8\u8ba4\u901a\u8fc7data\u83b7\u5f97\u7279\u5f81\u503cx\uff0c\u901a\u8fc7target\u83b7\u5f97\u76ee\u6807\u503cy\n<\/code><\/pre>\n<\/li>\n<li>\n<p>\u65b9\u5f0f\u4e00\uff1a\u7559\u51fa\u6cd5-\u968f\u673a\u5206\u5272<\/p>\n<pre><code class=\"language-python\">#2.\u7559\u51fa\u6cd5\uff08\u968f\u673a\u5206\u5272\uff09\nx_train,x_test,y_train,y_test = train_test_split(x,y,test_size=0.2) \n#test_size\u8868\u793a\u6d4b\u8bd5\u96c6\u6bd4\u4f8b\n<\/code><\/pre>\n<\/li>\n<li>\n<p>\u65b9\u5f0f\u4e8c\uff1a\u7559\u51fa\u6cd5-\u5206\u5c42\u5206\u5272<\/p>\n<pre><code class=\"language-python\">#2.\u7559\u51fa\u6cd5\uff08\u5206\u5c42\u5206\u5272\uff09\nx_train,x_test,y_train,y_test = train_test_split(x,y,test_size=0.2,stratify=y)\n#test_size\u8868\u793a\u6d4b\u8bd5\u96c6\u6bd4\u4f8b\n#stratify=y\u8868\u793a\u4fdd\u8bc1\u6d4b\u8bd5\u96c6\u548c\u8bad\u7ec3\u96c6\u4e4b\u95f4\u7684y\u6807\u7b7e\u4e0b\u7684\u6bd4\u4f8b\u76f8\u540c\n<\/code><\/pre>\n<\/li>\n<li>\n<p>\u65b9\u5f0f\u4e09\uff1a\u591a\u6b21\u5212\u5206-\u968f\u673a\u5206\u5272<\/p>\n<pre><code class=\"language-python\">#2.\u591a\u6b21\u5212\u5206\uff08\u968f\u673a\u5206\u5272\uff09\nspliter = ShuffleSplit(n_splits=5,test_size=0.2,random_state=0)\n#n_splits\uff1a\u8868\u793a\u5206\u6210\u51e0\u4e2a\u6570\u636e\u96c6\n#random_state\uff1a\u968f\u673a\u6570\u79cd\u5b50\n<\/code><\/pre>\n<\/li>\n<li>\n<p>\u65b9\u5f0f\u56db\uff1a\u591a\u6b21\u5212\u5206-\u5206\u5c42\u5206\u5272<\/p>\n<pre><code class=\"language-python\">#2.\u591a\u6b21\u5212\u5206\uff08\u5206\u5c42\u5206\u5272\uff09\nspliter = StratifiedShuffleSplit(n_splits=5,test_size=0.2,random_state=0)\n<\/code><\/pre>\n<\/li>\n<\/ul>\n<h4>4.\u4ea4\u53c9\u9a8c\u8bc1\u6cd5<\/h4>\n<p>K-Ford\u4ea4\u53c9\u9a8c\u8bc1\uff1a\u5c06\u8bad\u7ec3\u96c6\u5747\u5300\u5206\u6210k\u4efd\uff0c\u5047\u8bbe\u6bcf\u4efd\u6570\u636e\u7684\u6807\u53f7\u4e3a0~k-1\uff1a<\/p>\n<ul>\n<li>\u4ece0\u5f00\u59cb\uff0c\u6bcf\u6b21\u62bd\u51fa\u6807\u53f7\u4e3ai(i\u4ece0\u5230k-1)\u7684\u6570\u636e<\/li>\n<li>\u4f7f\u7528\u5269\u4e0b\u7684\u6570\u636e\u8fdb\u884c\u8bad\u7ec3\u5f62\u6210k\u4e2a\u6a21\u578b<\/li>\n<li>\u518d\u4f7f\u7528\u62bd\u51fa\u7684\u6570\u636e\u8fdb\u884c\u68c0\u9a8c<\/li>\n<\/ul>\n<pre><code class=\"language-python\">from sklearn.model_selection import KFold\nfrom sklearn.model_selection import StratifiedKFold\nfrom collections import Counter\nfrom sklearn.datasets import load_iris\n\n#1.\u52a0\u8f7d\u6570\u636e\u96c6\nx,y = load_iris(return_X_y=True)\nprint(&quot;\u539f\u59cb\u6bd4\u4f8b:&quot;,Counter(y))\n\n#2.\u968f\u673a\u4ea4\u53c9\u9a8c\u8bc1\nspliter = KFold(n_splits=5,shuffle=True,random_state=0)\nfor train,test in spliter.split(x,y):\n    print(&quot;\u968f\u673a\u4ea4\u53c9\u9a8c\u8bc1:&quot;,Counter(y[test]))\n        \n#3.\u5206\u5c42\u4ea4\u53c9\u9a8c\u8bc1\nspliter = StratifiedKFold(n_splits=5,shuffle=True,random_state=0)\nfor train,test in spliter.split(x,y):\n    print(&quot;\u5206\u5c42\u4ea4\u53c9\u9a8c\u8bc1&quot;,Counter(y[test]))\n<\/code><\/pre>\n<h4>5.\u7559\u4e00\u6cd5<\/h4>\n<pre><code class=\"language-python\">from sklearn.model_selection import LeaveOneOut\nfrom sklearn.model_selection import LeavePOut\nfrom sklearn.datasets import load_iris\nfrom collections import Counter\n\n#1.\u52a0\u8f7d\u6570\u636e\u96c6\nx,y = load_iris(return_X_y=True)\nprint(&quot;\u539f\u59cb\u6bd4\u4f8b:&quot;,Counter(y))\n\n#2.\u7559\u4e00\u6cd5\nspliter = LeaveOneOut()\nfor train,test in spliter.split(x,y):\n    print(&quot;\u8bad\u7ec3\u96c6:&quot;,len(train),&quot;\u6d4b\u8bd5\u96c6:&quot;,len(test))\n    \n#3.\u7559P\u6cd5\nspliter = LeavePOut(p=3) #\u8868\u793a\u6bcf\u6b21\u62bd\u53d6p\u4e2a\u6570\u636e\nfor train,test in spliter.split(x,y):\n    print(&quot;\u8bad\u7ec3\u96c6:&quot;,len(train),&quot;\u6d4b\u8bd5\u96c6:&quot;,len(test))\n<\/code><\/pre>\n<h4>6.\u81ea\u52a9\u6cd5<\/h4>\n<pre><code class=\"language-python\">import pandas as pd\n\n#1.\u6784\u9020\u6570\u636e\u96c6\ndata = [\u6570\u636e]\ndata = pd.DataFrame(data)\n\n#2.\u4ea7\u751f\u8bad\u7ec3\u96c6\ntrain = data.sample(frac=1,replace=True)\nprint(&quot;\u8bad\u7ec3\u96c6:&quot;,train)\n\n#3.\u4ea7\u751f\u6d4b\u8bd5\u96c6\ntest = data.loc[data.index.difference(train.index)]\nprint(&quot;\u6d4b\u8bd5\u96c6:&quot;,test)\n<\/code><\/pre>\n<h3>\uff08\u516d\uff09\u5206\u7c7b\u7b97\u6cd5\u7684\u8bc4\u4f30<\/h3>\n<h4>1.\u8bc4\u4f30\u65b9\u6cd5<\/h4>\n<ul>\n<li>\n<p>\u5229\u7528\u8bad\u7ec3\u597d\u7684\u6a21\u578b\u4f7f\u7528\u6d4b\u8bd5\u96c6\u7684\u7279\u5f81\u503c\u8fdb\u884c\u9884\u6d4b<\/p>\n<\/li>\n<li>\n<p>\u5c06\u9884\u6d4b\u7ed3\u679c\u548c\u6d4b\u8bd5\u96c6\u7684\u76ee\u6807\u503c\u8fdb\u884c\u6bd4\u8f83\uff0c\u8ba1\u7b97\u9884\u6d4b\u7684\u51c6\u786e\u7387<\/p>\n<pre><code class=\"language-python\">from sklearn import datasets\nfrom sklearn.model_selection import train_test_split\nfrom sklearn.neighbors import KNeighborsClassifier\n\n#1.\u52a0\u8f7d\u6570\u636e\u96c6\nx,y = datasets.load_iris(return_X_y=True)\n\n#2.\u5212\u5206\u8bad\u7ec3\u96c6\u548c\u6d4b\u8bd5\u96c6\nx_train,x_test,y_train,y_test = train_test_split(x,y,test_size=0.2)\n\n#3.\u521b\u5efaKNN\u5bf9\u8c61\nestimator = KNeighborsClassifier(n_neighbors=6)\n\n#4.\u4f7f\u7528KNN\u8bad\u7ec3\u6a21\u578b\nestimator.fit(x,y)\n\n#5.\u8fdb\u884c\u9884\u6d4b\npredict = estimator.predict(x_test)\n\n#6.\u8fdb\u884c\u8bc4\u4f30\nprint(sum(predict==y_test)\/y_test.shape[0])\n<\/code><\/pre>\n<\/li>\n<\/ul>\n<h4>2.\u8bc4\u4f30API<\/h4>\n<pre><code class=\"language-python\">from sklearn.metrics import accuracy_score\n\n#\u65b9\u5f0f1\uff1a\u76f4\u63a5\u8c03\u7528\u51fd\u6570\naccuracy_score(\u9884\u6d4b\u503c,\u6d4b\u8bd5\u503c)\n\n#\u65b9\u5f0f2\uff1a\u901a\u8fc7\u5bf9\u8c61\u8c03\u7528\u51fd\u6570\nestimator.score(\u6d4b\u8bd5\u96c6\u7279\u5f81\u503c,\u6d4b\u8bd5\u96c6\u76ee\u6807\u503c)\n<\/code><\/pre>\n<h3>\uff08\u4e03\uff09K\u503c\u9009\u62e9<\/h3>\n<p>KNN\u7b97\u6cd5\u7684\u6838\u5fc3\u5728\u4e8e\u627e\u5230\u5408\u9002\u7684K\u503c<\/p>\n<h4>1.\u5982\u4f55\u786e\u5b9aK\u503c<\/h4>\n<ul>\n<li>K\u503c\u8fc7\u5c0f\uff1a\u5bb9\u6613\u53d7\u5230\u5f02\u5e38\u70b9\u7684\u5f71\u54cd<\/li>\n<li>K\u503c\u8fc7\u5927\uff1a\u5bb9\u6613\u53d7\u5230\u6837\u672c\u5747\u8861\u7684\u5f71\u54cd<\/li>\n<li>k=\u6837\u672c\u6570\uff1a\u5bfc\u81f4\u7ed3\u679c\u6c38\u8fdc\u662f\u6837\u672c\u4e2d\u591a\u7684\u90a3\u4e00\u4e2a<\/li>\n<\/ul>\n<p>\u56e0\u6b64\uff0ck\u4e00\u822c\u53d6\u4e00\u4e2a\u8f83\u5c0f\u7684\u6570\u503c\uff0c\u901a\u8fc7\u4ea4\u53c9\u9a8c\u8bc1\u6cd5\u6765\u9009\u62e9\u6700\u4f18\u7684k\u503c<\/p>\n<h4>2.GridSearchCV<\/h4>\n<pre><code class=\"language-python\">from sklearn.datasets import load_iris\nfrom sklearn.model_selection import train_test_split\nfrom sklearn.neighbors import KNeighborsClassifier\nfrom sklearn.model_selection import GridSearchCV\n\n#1.\u52a0\u8f7d\u6570\u636e\u96c6\nx,y = load_iris(return_X_y=True)\nx_train,x_test,y_train,y_test = train_test_split(x,y,test_size=0.2,stratify=y,random_state=0)\n\n#2.\u5bfb\u627e\u6700\u4f18\u53c2\u6570\nestimator = KNeighborsClassifier()\nparam_grid = {&quot;n_neighbors&quot;:[1,2,3,4,5,6,7,8,9,10]} #\u4f7f\u7528\u5b57\u5178\u786e\u5b9a\u8981\u4f20\u7684\u53c2\u6570\u4ee5\u53ca\u53ef\u80fd\u7684\u53d6\u503c\nestimator = GridSearchCV(estimator,param_grid=param_grid,cv=5,verbose=0)\n#cv\uff1a\u8868\u793a\u5c06\u6570\u636e\u96c6\u5212\u5206\u4e3a5\u4efd\n#verbose\uff1a\u662f\u5426\u663e\u793a\u6253\u5370\u4fe1\u606f\nestimator.fit(x_train,y_train)\n\n#3.\u8f93\u51fa\u7ed3\u679c\nprint(&quot;\u6700\u4f18\u53c2\u6570:&quot;,estimator.best_params_,&quot;\u6700\u4f18\u5f97\u5206\uff1a&quot;,estimator.best_score_) # \u6ce8\u610f\uff1a\u6700\u4f18\u5f97\u5206\u662f\u9a8c\u8bc1\u96c6\u7684\u7ed3\u679c\nprint(&quot;\u51c6\u786e\u7387:&quot;,estimator.score(x_test,y_test)) #\u51c6\u786e\u7387\u662f\u6d4b\u8bd5\u96c6\u7684\u7ed3\u679c\n<\/code><\/pre>\n<h2>\u4e94\u3001\u7ebf\u6027\u56de\u5f52\u7b97\u6cd5<\/h2>\n<h3>\uff08\u4e00\uff09\u7ebf\u6027\u56de\u5f52\u7b97\u6cd5\u7b80\u4ecb<\/h3>\n<h4>1.\u6982\u5ff5<\/h4>\n<p>\u7ebf\u6027\u56de\u5f52\uff1a\u662f\u6307\u5229\u7528\u56de\u5f52\u65b9\u7a0b\u5bf9\u4e00\u4e2a\u6216\u591a\u4e2a\u81ea\u53d8\u91cf\u548c\u56e0\u53d8\u91cf\u4e4b\u95f4\u7684\u5173\u7cfb\u8fdb\u884c\u5efa\u6a21\u7684\u4e00\u79cd\u5206\u6790\u65b9\u5f0f<\/p>\n<h4>2.\u516c\u5f0f<\/h4>\n<p>$$\nh(w)=w_1x_1+w_2x_2+w_3x_3+\u2026+b=w^Tx+b\n$$<\/p>\n<h4>3.\u7ebf\u6027\u56de\u5f52API<\/h4>\n<ul>\n<li>\u6a21\u5757\uff1alinear_model<\/li>\n<li>\u5bf9\u8c61\uff1asklearn.linear_model.LinearRegression()\n<ul>\n<li>\u53c2\u6570\n<ul>\n<li>fit_intercept\uff1a\u662f\u5426\u8ba1\u7b97\u504f\u7f6e<\/li>\n<\/ul>\n<\/li>\n<li>\u5c5e\u6027\n<ul>\n<li>LinearRegression.coef_\uff1a\u56de\u5f52\u7cfb\u6570<\/li>\n<li>LinearRegression.intercept_\uff1a\u504f\u7f6e<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n<h4>4.\u57fa\u672c\u6d41\u7a0b<\/h4>\n<pre><code class=\"language-python\">from sklearn.linear_model import LinearRegression\n\ntrain = \u8bad\u7ec3\u96c6\ntest = \u6d4b\u8bd5\u96c6\n#\u5b9e\u4f8b\u5316\u7ebf\u6027\u56de\u5f52\u5bf9\u8c61\nestimator = LinearRegression()\n#\u8bad\u7ec3\u6a21\u578b\nestimator.fit(train,test)\n#\u67e5\u770b\u56de\u5f52\u7cfb\u6570\nprint(estimator.coef_)\n#\u8fdb\u884c\u9884\u6d4b\nestimator.predict(\u9884\u6d4b)\n<\/code><\/pre>\n<h3>\uff08\u4e8c\uff09\u635f\u5931\u51fd\u6570<\/h3>\n<h4>1.\u635f\u5931\u51fd\u6570\u7684\u6982\u5ff5<\/h4>\n<p>\u635f\u5931\u51fd\u6570\uff1a\u7528\u4e8e\u8ba1\u7b97\u9884\u6d4b\u503c\u548c\u771f\u5b9e\u503c\u4e4b\u95f4\u7684\u8bef\u5dee\uff0c\u8bef\u5dee\u8d8a\u5c0f\u8bf4\u660e\u6a21\u578b\u6027\u80fd\u8d8a\u597d<\/p>\n<h4>2.\u5e73\u65b9\u635f\u5931<\/h4>\n<p>$$\nJ(w)=(h(x_1)-y_1)^2+(h(x_2)-y_2)^2+\u2026+(h(x_n)-y_n)^2=\\sum_{i=1}^n (h(x_i)-y_i)^2\n$$<\/p>\n<h4>3.\u6b63\u89c4\u65b9\u7a0b<\/h4>\n<p>\u5bf9\u635f\u5931\u51fd\u6570\u6c42\u5bfc\u7b49\u96f6\uff0c\u8ba1\u7b97\u6700\u503c\n$$\nw=(X^TX)^{-1}X^Ty\n$$<\/p>\n<h3>\uff08\u4e09\uff09\u68af\u5ea6\u4e0b\u964d<\/h3>\n<h4>1.\u7b80\u4ecb<\/h4>\n<p>\u5bf9\u4e8e\u4e00\u4e2a\u53ef\u5fae\u5206\u7684\u51fd\u6570\uff0c\u5bfb\u627e\u5176\u6700\u503c\uff1a<\/p>\n<ul>\n<li>\u627e\u5230\u5f53\u524d\u4f4d\u7f6e\u7684\u6700\u5927\u68af\u5ea6<\/li>\n<li>\u6cbf\u7740\u68af\u5ea6\u7684\u6b63\u65b9\u5411\uff08\u6c42\u6700\u5927\u503c\uff09\u6216\u8005\u53cd\u65b9\u5411\uff08\u6c42\u6700\u5c0f\u503c\uff09<\/li>\n<li>\u91cd\u590d\u4e0a\u8ff0\u6b65\u9aa4\uff0c\u5c31\u80fd\u4e0d\u65ad\u903c\u8fd1\u5c40\u90e8\u6700\u5c0f\u503c<\/li>\n<\/ul>\n<h4>2.\u516c\u5f0f<\/h4>\n<p>$$\n\\theta_{i+1}=\\theta_{i}-\\alpha\\frac{\\partial}{\\partial \\theta_{i}}J(\\theta)\n$$<\/p>\n<ul>\n<li>\n<p>$ \\alpha $\uff1a\u8868\u793a\u5b66\u4e60\u7387\uff0c\u4e5f\u79f0\u4e3a\u6b65\u957f\uff0c\u63a7\u5236\u6bcf\u4e00\u6b65\u8d70\u7684\u8ddd\u79bb<\/p>\n<ul>\n<li>\u6ce8\u610f\uff1a\u5982\u679c\u8fc7\u5927\uff0c\u5219\u53ef\u80fd\u9519\u8fc7\u6700\u503c\u70b9\uff1b\u5982\u679c\u8fc7\u5c0f\uff0c\u5219\u53ef\u80fd\u9700\u8981\u591a\u6b21\u8fed\u4ee3<\/li>\n<\/ul>\n<\/li>\n<li>\n<p>$ &#8211; $\uff1a\u8d1f\u53f7\u8868\u793a\u6cbf\u7740\u68af\u5ea6\u4e0b\u964d\u6700\u5feb\u7684\u65b9\u5411\u79fb\u52a8\uff0c\u56e0\u4e3a\u68af\u5ea6\u672c\u8eab\u8868\u793a\u7684\u662f\u4e0a\u5347\u6700\u5feb\u7684<\/p>\n<\/li>\n<\/ul>\n<h4>3.\u68af\u5ea6\u4e0b\u964d\u6d41\u7a0b<\/h4>\n<ul>\n<li>\u524d\u63d0\uff1a\u786e\u5b9a\u4f18\u5316\u6a21\u578b\u7684\u5047\u8bbe\u51fd\u6570\u548c\u635f\u5931\u51fd\u6570<\/li>\n<li>\u7b97\u6cd5\u76f8\u5173\u53c2\u6570\u521d\u59cb\u5316<\/li>\n<li>\u901a\u8fc7\u516c\u5f0f\u8fdb\u884c\u8fed\u4ee3<\/li>\n<\/ul>\n<h3>\uff08\u56db\uff09\u5176\u4ed6\u68af\u5ea6\u4e0b\u964d\u7b97\u6cd5<\/h3>\n<h4>1.\u5168\u68af\u5ea6\u4e0b\u964d\u7b97\u6cd5\uff08FGD\uff09<\/h4>\n<ul>\n<li>\n<p>\u5168\u68af\u5ea6\u4e0b\u964d\u7b97\u6cd5\uff1a\u5728\u6bcf\u6b21\u8fed\u4ee3\u65f6\uff0c\u4f7f\u7528\u5168\u90e8\u6837\u672c\u7684\u68af\u5ea6\u503c<\/p>\n<\/li>\n<li>\n<p>\u5177\u4f53\u5b9e\u73b0\uff1a<\/p>\n<ul>\n<li>\u5728\u66f4\u65b0\u53c2\u6570\u65f6\uff0c\u4f7f\u7528\u6240\u6709\u7684\u6837\u672c\u6765\u8fdb\u884c\u66f4\u65b0<\/li>\n<li>\u8ba1\u7b97\u8bad\u7ec3\u96c6\u7684\u6240\u6709\u6837\u672c\u8bef\u5dee\uff0c\u5bf9\u5176\u6c42\u548c\u5e73\u5747\u540e\u4f5c\u4e3a\u76ee\u6807\u51fd\u6570<\/li>\n<\/ul>\n<\/li>\n<li>\n<p>\u516c\u5f0f\uff1a\n$$\n\\theta_{i+1}=\\theta_{i}-\\alpha\\sum_{j=0}^{m}(h_{\\theta}(x_0^{(j)},x_2^{(j)},\u2026,x_n^{(j)})-y_i)x_i^{(j)}\n$$<\/p>\n<\/li>\n<li>\n<p>\u6ce8\u610f\uff1a<\/p>\n<ul>\n<li>\u5168\u68af\u5ea6\u4e0b\u964d\u7b97\u6cd5\u6267\u884c\u901f\u5ea6\u6162\uff0c\u65e0\u6cd5\u5904\u7406\u8d85\u51fa\u5185\u5b58\u5bb9\u91cf\u9650\u5236\u7684\u6570\u636e\u96c6<\/li>\n<li>\u5168\u68af\u5ea6\u4e0b\u964d\u7b97\u6cd5\u4e0d\u652f\u6301\u5728\u7ebf\u66f4\u65b0\u6a21\u578b\uff0c\u5728\u8fd0\u884c\u8fc7\u7a0b\u4e2d\uff0c\u4e0d\u80fd\u589e\u52a0\u65b0\u7684\u6837\u672c<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n<h4>2.\u968f\u673a\u68af\u5ea6\u4e0b\u964d\u7b97\u6cd5\uff08SGD\uff09<\/h4>\n<ul>\n<li>\n<p>\u968f\u673a\u68af\u5ea6\u4e0b\u964d\u7b97\u6cd5\uff1a\u5728\u6bcf\u6b21\u8fed\u4ee3\u65f6\uff0c\u968f\u673a\u9009\u53d6\u4e00\u4e2a\u6837\u672c\u68af\u5ea6\u503c<\/p>\n<\/li>\n<li>\n<p>\u516c\u5f0f\uff1a\n$$\n\\theta_{i+1}=\\theta_{i}-\\alpha(h_{\\theta}(x_0^{(j)},x_2^{(j)},\u2026,x_n^{(j)})-y_i)x_i^{(j)}\n$$<\/p>\n<\/li>\n<li>\n<p>\u6ce8\u610f\uff1a<\/p>\n<ul>\n<li>\u968f\u673a\u68af\u5ea6\u4e0b\u964d\u7b97\u6cd5\u5728\u906d\u9047\u566a\u58f0\u65f6\u5bb9\u6613\u9677\u5165\u5c40\u90e8\u6700\u4f18\u89e3<\/li>\n<\/ul>\n<\/li>\n<li>\n<p>API\uff1a<code>from sklearn.linear_model import SGDRegresser<\/code><\/p>\n<ul>\n<li>\u53c2\u6570\n<ul>\n<li>loss\uff1a\u635f\u5931\u51fd\u6570\u7c7b\u578b<\/li>\n<li>fit_intercept\uff1a\u662f\u5426\u8ba1\u7b97\u504f\u7f6e<\/li>\n<li>learning_rate\uff1a\u5b66\u4e60\u7387<\/li>\n<\/ul>\n<\/li>\n<li>\u5c5e\u6027\n<ul>\n<li>SGDRegression.coef_\uff1a\u56de\u5f52\u7cfb\u6570<\/li>\n<li>SGDRegression.intercept_\uff1a\u504f\u7f6e<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n<h4>3.\u5c0f\u6279\u91cf\u68af\u5ea6\u4e0b\u964d\u7b97\u6cd5<\/h4>\n<ul>\n<li>\n<p>\u5c0f\u6279\u91cf\u68af\u5ea6\u4e0b\u964d\u7b97\u6cd5\uff1a\u5728\u6bcf\u6b21\u8fed\u4ee3\u65f6\uff0c\u968f\u673a\u62bd\u53d6\u4e00\u4e2a\u5c0f\u7684\u6837\u672c\u96c6\uff08\u901a\u5e38\u4e3a2\u7684\u6b21\u5e42\uff0c\u4fbf\u4e8eGPU\u8fd0\u7b97\uff09\u4f7f\u7528\u5168\u68af\u5ea6\u4e0b\u964d\u7b97\u6cd5<\/p>\n<\/li>\n<li>\n<p>\u516c\u5f0f\uff1a\n$$\n\\theta_{i+1}=\\theta_{i}-\\alpha\\sum_{j=t}^{t+x-1}(h_{\\theta}(x_0^{(j)},x_2^{(j)},\u2026,x_n^{(j)})-y_i)x_i^{(j)}\n$$<\/p>\n<\/li>\n<\/ul>\n<h4>4.\u968f\u673a\u5e73\u5747\u68af\u5ea6\u4e0b\u964d\u7b97\u6cd5<\/h4>\n<ul>\n<li>\n<p>\u968f\u673a\u5e73\u5747\u68af\u5ea6\u4e0b\u964d\u7b97\u6cd5\uff1a\u5728\u5185\u5b58\u4e2d\u4e3a\u6240\u6709\u6837\u672c\u7ef4\u62a4\u4e00\u4e2a\u65e7\u68af\u5ea6\u503c\uff0c\u5728\u6bcf\u6b21\u8fed\u4ee3\u65f6\uff0c\u968f\u673a\u9009\u62e9\u4e00\u4e2a\u65b0\u7684\u6837\u672c\u6765\u66f4\u65b0\u6b64\u6837\u672c\u7684\u68af\u5ea6\uff0c\u5176\u4ed6\u6837\u672c\u7684\u68af\u5ea6\u4fdd\u6301\u4e0d\u53d8\uff0c\u7136\u540e\u6c42\u5e73\u5747\u68af\u5ea6\u66f4\u65b0\u53c2\u6570<\/p>\n<\/li>\n<li>\n<p>\u516c\u5f0f\uff1a\n$$\n\\theta_{i+1}=\\theta_{i}-\\frac{\\alpha}{n}(h_{\\theta}(x_0^{(j)},x_2^{(j)},\u2026,x_n^{(j)})-y_i)x_i^{(j)}\n$$<\/p>\n<\/li>\n<\/ul>\n<h3>\uff08\u4e94\uff09\u56de\u5f52\u95ee\u9898\u8bc4\u4f30<\/h3>\n<h4>1.\u5e73\u5747\u7edd\u5bf9\u8bef\u5dee\uff08MAE\uff09<\/h4>\n<ul>\n<li>\n<p>\u516c\u5f0f\uff1a\n$$\nMAE=\\frac{1}{n}\\sum_{i=1}^{n} |y_{i}-\\hat y_{i}|\n$$<\/p>\n<\/li>\n<li>\n<p>API<\/p>\n<pre><code class=\"language-python\">from sklearn.metrics import mean_absolute_error\nmean_absolute_error(\u9884\u6d4b\u503c,\u771f\u5b9e\u503c)\n<\/code><\/pre>\n<\/li>\n<\/ul>\n<h4>2.\u5747\u65b9\u8bef\u5dee\uff08MSE\uff09<\/h4>\n<ul>\n<li>\n<p>\u516c\u5f0f\uff1a\n$$\nMSE=\\frac{1}{n}\\sum_{i=1}^{n} (y_{i}-\\hat y_{i})^2\n$$<\/p>\n<\/li>\n<li>\n<p>API\uff1a<\/p>\n<pre><code class=\"language-python\">from sklearn.metrics import mean_squared_error\nmean_squared_error(\u9884\u6d4b\u503c,\u771f\u5b9e\u503c)\n<\/code><\/pre>\n<\/li>\n<\/ul>\n<h4>3.\u5747\u65b9\u6839\u8bef\u5dee\uff08RMSE\uff09<\/h4>\n<ul>\n<li>\u516c\u5f0f\uff1a\n$$\nMSE=\\sqrt{\\frac{1}{n}\\sum_{i=1}^{n} (y_{i}-\\hat y_{i})^2}\n$$<\/li>\n<\/ul>\n<h4>4.R-Squared<\/h4>\n<ul>\n<li>\n<p>\u516c\u5f0f\uff1a\n$$\nR^2=1-\\frac{\\sum(y_i-\\hat y_i)^2}{\\sum(y_i-\\overline y_i)^2}\n$$<\/p>\n<\/li>\n<li>\n<p>API\uff1a<\/p>\n<pre><code class=\"language-python\">from sklearn.metrics import r2_score\nr2_score(\u9884\u6d4b\u503c,\u771f\u5b9e\u503c)\n<\/code><\/pre>\n<\/li>\n<li>\n<p>\u6ce8\u610f\uff1a<\/p>\n<ul>\n<li>\u4e00\u822c\u6765\u8bf4\uff0c\u6211\u4eec\u8ba4\u4e3a0.5\u4ee5\u4e0b\u4e3a\u5f31\u62df\u5408\uff0c0.5~0.8\u4e3a\u4e2d\u62df\u5408\uff0c0.8\u4ee5\u4e0a\u4e3a\u5f3a\u62df\u5408\uff08\u7406\u8bba\u4e0a\uff0c\u5c0f\u4e8e0\u7684\u60c5\u51b5\u662f\u5b58\u5728\u7684\uff0c\u8868\u793a\u6a21\u578b\u975e\u5e38\u5dee\uff09<\/li>\n<li>R-Squared\u53ea\u662f\u8868\u793a\u62df\u5408\u7684\u5f3a\u5f31\uff0c\u5e76\u4e0d\u80fd\u8ba4\u4e3aR-Squared\u8d8a\u5927\u6a21\u578b\u8d8a\u597d<\/li>\n<li>R-Squared\u5bf9\u975e\u7ebf\u6027\u6a21\u578b\u6ca1\u6709\u610f\u4e49<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n<h4>5.Adjusted R-Squared<\/h4>\n<ul>\n<li>\n<p>\u516c\u5f0f\uff1a\n$$\nR_{adj}^2=1-\\frac{(1-R^2)(n-1)}{n-k-1}\n$$<\/p>\n<ul>\n<li>n\uff1a\u6837\u672c\u6570\u91cf<\/li>\n<li>k\uff1a\u6a21\u578b\u4e2d\u81ea\u53d8\u91cf\u7684\u6570\u91cf<\/li>\n<\/ul>\n<\/li>\n<li>\n<p>\u6ce8\u610f\uff1a<\/p>\n<ul>\n<li>\u5982\u679cAdjust R-Squared\u548cR-Squared\u5dee\u8ddd\u8f83\u5927\uff0c\u5219\u8bf4\u660e\u6a21\u578b\u5b58\u5728\u8fc7\u62df\u5408<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n<h3>\uff08\u516d\uff09\u6b63\u5219\u5316<\/h3>\n<h4>1.\u6982\u5ff5<\/h4>\n<p>\u6b63\u5219\u5316\uff1a\u5728\u7b97\u6cd5\u5b66\u4e60\u8fc7\u7a0b\u4e2d\uff0c\u6570\u636e\u63d0\u4f9b\u7684\u67d0\u4e9b\u7279\u5f81\u4e2d\u53ef\u80fd\u5f71\u54cd\u6a21\u578b\u590d\u6742\u5ea6\uff0c\u6216\u8005\u8fd9\u4e2a\u7279\u5f81\u7684\u5f02\u5e38\u8f83\u591a\uff0c\u6211\u4eec\u9700\u8981\u51cf\u5c11\u8fd9\u4e2a\u7279\u5f81\u7684\u5f71\u54cd\uff0c\u5c31\u662f\u6b63\u5219\u5316<\/p>\n<h4>2.\u7c7b\u522b<\/h4>\n<ul>\n<li>\n<p>L1\u6b63\u5219\u5316<\/p>\n<ul>\n<li>\n<p>\u516c\u5f0f\uff1a\u5047\u8bbeL(w)\u662f\u672a\u6dfb\u52a0\u6b63\u5219\u5316\u9879\u7684\u635f\u5931\u51fd\u6570\uff0c$ \\lambda $\u63a7\u5236\u6b63\u5219\u5316\u9879\u7684\u5927\u5c0f\n$$\n\u6700\u7ec8\u635f\u5931\u51fd\u6570\uff1aL=L(w)+\\lambda*\\sum_{i=1}^n |w_i|\n$$<\/p>\n<\/li>\n<li>\n<p>\u4f5c\u7528\uff1a\u8fdb\u884c\u7279\u5f81\u9009\u62e9\uff0cL1\u6b63\u5219\u5316\u4f1a\u4ea7\u751f\u8f83\u591a\u7684\u53c2\u6570\u4e3a0\uff0c\u4ece\u800c\u4ea7\u751f\u7a00\u758f\u89e3\uff0c\u5c06\u53c2\u6570\u4e3a0\u7684\u7279\u5f81\u820d\u5f03\uff0c\u5373\u53ef\u8fbe\u5230\u7279\u5f81\u9009\u62e9\u7684\u76ee\u7684<\/p>\n<\/li>\n<li>\n<p>API\uff1a<code>form sklearn.linear_model import Lasso<\/code><\/p>\n<ul>\n<li>alpha<\/li>\n<li>normalize\uff1aTrue\/False\uff0c\u662f\u5426\u6807\u51c6\u5316<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n<\/li>\n<li>\n<p>L2\u6b63\u5219\u5316<\/p>\n<ul>\n<li>\n<p>\u516c\u5f0f\uff1a\u5047\u8bbeL(w)\u662f\u672a\u6dfb\u52a0\u6b63\u5219\u5316\u9879\u7684\u635f\u5931\u51fd\u6570\uff0c$ \\lambda $\u63a7\u5236\u6b63\u5219\u5316\u9879\u7684\u5927\u5c0f\n$$\n\u6700\u7ec8\u635f\u5931\u51fd\u6570\uff1aL=L(w)+\\lambda*\\sum_{i=1}^n w_i^2\n$$<\/p>\n<\/li>\n<li>\n<p>\u4f5c\u7528\uff1a\u9632\u6b62\u6a21\u578b\u8fc7\u62df\u5408\uff0c\u51cf\u5c0f\u7279\u5f81\u7684\u6743\u91cd<\/p>\n<\/li>\n<li>\n<p>API\uff1a<code>form sklearn.linear_model import Ridge<\/code><\/p>\n<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n<h2>\u516d\u3001\u903b\u8f91\u56de\u5f52\u7b97\u6cd5<\/h2>\n<h3>\uff08\u4e00\uff09\u903b\u8f91\u56de\u5f52\u5e94\u7528\u573a\u666f<\/h3>\n<p>\u903b\u8f91\u56de\u5f52\uff1a\u662f\u673a\u5668\u5b66\u4e60\u4e2d\u4e00\u79cd\u5206\u7c7b\u6a21\u578b\uff0c\u7528\u4e8e\u89e3\u51b3\u4e8c\u5206\u7c7b\u95ee\u9898<\/p>\n<h3>\uff08\u4e8c\uff09\u903b\u8f91\u56de\u5f52\u7684\u539f\u7406<\/h3>\n<h4>1.\u8f93\u5165<\/h4>\n<p>\u903b\u8f91\u56de\u5f52\u7684\u8f93\u5165\u662f\u4e00\u4e2a\u7ebf\u6027\u65b9\u7a0b\n$$\nh(w)=w_1x_1+w_2x_2+w_3x_3+\u2026+b\n$$<\/p>\n<h4>2.\u6fc0\u6d3b\u51fd\u6570<\/h4>\n<p>sigmoid\u51fd\u6570\uff1a\n$$\ng(w^T,x)=\\frac{1}{1+e^{-h(w)}}=\\frac{1}{1+e^{-w^Tx}}\n$$\n\u56de\u5f52\u7ed3\u679c\u8f93\u5165\u5230sigmoid\u51fd\u6570\u4e2d\uff0c\u8f93\u51fa[0,1]\u7684\u4e00\u4e2a\u6982\u7387\u503c\uff0c\u9ed8\u8ba4\u9608\u503c\u4e3a0.5<\/p>\n<h3>\uff08\u4e09\uff09\u635f\u5931\u4ee5\u53ca\u4f18\u5316<\/h3>\n<h4>1.\u635f\u5931<\/h4>\n<p>\u903b\u8f91\u56de\u5f52\u7684\u635f\u5931\uff0c\u79f0\u4e4b\u4e3a\u5bf9\u6570\u4f3c\u7136\u635f\u5931<\/p>\n<ul>\n<li>\u5206\u7c7b\u5f62\u5f0f\u7684\u516c\u5f0f\u4e3a\uff1a<\/li>\n<\/ul>\n<p>$$\ncost(h_\\theta(x),y)=\\begin{cases}-log(h_\\theta(x))\uff0c\u5f53y=1\u65f6\\\\-log(1-h_\\theta(x))\uff0c\u5f53y=0\u65f6\\end{cases}\n$$<\/p>\n<ul>\n<li>\u7efc\u5408\u5b8c\u6574\u635f\u5931\u51fd\u6570\uff1a\n$$\ncost(h_\\theta(x),y)=\\sum_{i=1}^m -y_i log(h_\\theta(x))-(1-y_i)log(1-h_\\theta(x))\n$$<\/li>\n<\/ul>\n<h4>2.\u4f18\u5316<\/h4>\n<p>\u4f7f\u7528\u68af\u5ea6\u4e0b\u964d\u7b97\u6cd5\u51cf\u5c11\u635f\u5931\u51fd\u6570\u7684\u503c\uff0c\u4ece\u800c\u66f4\u65b0\u903b\u8f91\u56de\u5f52\u5bf9\u5e94\u7b97\u6cd5\u7684\u6743\u91cd\u53c2\u6570<\/p>\n<h3>\uff08\u56db\uff09\u903b\u8f91\u56de\u5f52\u7684API<\/h3>\n<p><code>sklearn.linear_model.LogisticRegression(solver=,penalty=,C=)<\/code><\/p>\n<ul>\n<li><code>solver<\/code>\uff1a\u9ed8\u8ba4<code>'liblinear'<\/code>\uff0c\u53ef\u9009<code>'sag'<\/code>\u3001<code>'saga'<\/code>\u3001<code>'newton-cg'<\/code>\u3001<code>'lbfgs'<\/code>\n<ul>\n<li><code>'liblinear'<\/code>\uff1a\u9002\u7528\u4e8e\u5c0f\u578b\u6570\u636e\u96c6<\/li>\n<li><code>'sag'<\/code>\u3001<code>'saga'<\/code>\uff1a\u9002\u7528\u4e8e\u5927\u578b\u6570\u636e\u96c6<\/li>\n<li>\u591a\u5206\u7c7b\u95ee\u9898\uff1a<code>'sag'<\/code>\u3001<code>'saga'<\/code>\u3001<code>'newton-cg'<\/code>\u3001<code>'lbfgs'<\/code><\/li>\n<\/ul>\n<\/li>\n<li><code>penalty<\/code>\uff1a\u6b63\u5219\u5316\u7684\u79cd\u7c7b<\/li>\n<li><code>C<\/code>\uff1a\u6b63\u5219\u5316\u529b\u5ea6<\/li>\n<\/ul>\n<h3>\uff08\u4e94\uff09\u5206\u7c7b\u56de\u5f52\u6307\u6807<\/h3>\n<h4>1.\u6df7\u6dc6\u77e9\u9635<\/h4>\n<table>\n<thead>\n<tr>\n<th style=\"text-align:center\"><\/th>\n<th style=\"text-align:center\">\u6b63\u4f8b<\/th>\n<th style=\"text-align:center\">\u5047\u4f8b<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td style=\"text-align:center\">\u6b63\u4f8b<\/td>\n<td style=\"text-align:center\">\u771f\u6b63\u4f8bTP<\/td>\n<td style=\"text-align:center\">\u4f2a\u53cd\u4f8bFN<\/td>\n<\/tr>\n<tr>\n<td style=\"text-align:center\">\u5047\u4f8b<\/td>\n<td style=\"text-align:center\">\u4f2a\u6b63\u4f8bFP<\/td>\n<td style=\"text-align:center\">\u771f\u53cd\u4f8bTN<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<p>\u6ce8\uff1a\u5de6\u4fa7\uff1a\u771f\u5b9e\u7ed3\u679c\u2013\u4e0a\u65b9\uff1a\u9884\u6d4b\u7ed3\u679c<\/p>\n<h4>2.\u51c6\u786e\u7387<\/h4>\n<p>\u51c6\u786e\u7387\uff1a\u53c8\u79f0\u67e5\u51c6\u7387\uff0c\u6307\u7684\u662f\u5bf9\u6b63\u4f8b\u6837\u672c\u7684\u9884\u6d4b\u51c6\u786e\u7387\n$$\nP=\\frac{TP}{TP+FP}\n$$<\/p>\n<h4>3.\u53ec\u56de\u7387<\/h4>\n<p>\u53ec\u56de\u7387\uff1a\u53c8\u79f0\u67e5\u5168\u7387\uff0c\u6307\u7684\u662f\u9884\u6d4b\u4e3a\u771f\u6b63\u4f8b\u6837\u672c\u5360\u6240\u6709\u771f\u6b63\u4f8b\u6837\u672c\u7684\u6bd4\u91cd\n$$\nR=\\frac{TP}{TP+FN}\n$$<\/p>\n<h4>4.ROC\u66f2\u7ebf\u4e0eAUC\u503c<\/h4>\n<ul>\n<li>\n<p>ROC\u66f2\u7ebf\uff1a\u6211\u4eec\u5206\u522b\u8003\u8651\u6b63\u8d1f\u6837\u672c\u7684\u60c5\u51b5<\/p>\n<ul>\n<li>\n<p>\u6b63\u6837\u672c\u4e2d\u88ab\u9884\u6d4b\u4e3a\u6b63\u6837\u672c\u7684\u6982\u7387\uff08TPR\uff09<\/p>\n<\/li>\n<li>\n<p>\u8d1f\u6837\u672c\u4e2d\u88ab\u9884\u6d4b\u4e3a\u6b63\u6837\u672c\u7684\u6982\u7387\uff08FPR\uff09<\/p>\n<\/li>\n<\/ul>\n<\/li>\n<li>\n<p>AUC\u503c\uff1a\u7531\u4e8eROC\u66f2\u7ebf\u56fe\u50cf\u8d8a\u9760\u8fd1(0,1)\u70b9\u6a21\u578b\u5bf9\u4e8e\u6b63\u8d1f\u6837\u672c\u7684\u8fa8\u522b\u80fd\u529b\u8d8a\u5f3a\uff0c\u800c\u56fe\u50cf\u8d8a\u9760\u8fd1(0,1)\u70b9\uff0cROC\u66f2\u7ebf\u4e0b\u9762\u79ef\u8d8a\u5927\uff0c\u6545\u6211\u4eec\u5c06ROC\u66f2\u7ebf\u4e0b\u9762\u79ef\u79f0\u4e3aAUC\u503c<\/p>\n<\/li>\n<li>\n<p>API\uff1a<\/p>\n<ul>\n<li>\n<p>\u5206\u7c7b\u8bc4\u4f30\u62a5\u544a\uff1a<\/p>\n<pre><code class=\"language-python\">sklearn.metrics.classification_report(y_true,y_pred,labels=[],target_names=None)\n\n&quot;&quot;&quot;\ny_true\uff1a\u771f\u5b9e\u7684\u76ee\u6807\u503c\ny_pred\uff1a\u9884\u6d4b\u7684\u76ee\u6807\u503c\nlabels\uff1a\u6307\u5b9a\u7c7b\u522b\u5bf9\u5e94\u7684\u6570\u5b57\ntarget_names\uff1a\u76ee\u6807\u7c7b\u522b\u7684\u540d\u79f0\n\nreturn\uff1a\u6bcf\u4e2a\u7c7b\u522b\u7684\u7cbe\u786e\u7387\u4e0e\u53ec\u56de\u7387\n&quot;&quot;&quot;\n<\/code><\/pre>\n<\/li>\n<li>\n<p>AUC\u8ba1\u7b97\uff1a<\/p>\n<pre><code class=\"language-python\">from sklearn.metrics import roc_auc_score\n\nsklearn.metrics.roc_auc_score(y_true,t_score)\n\n&quot;&quot;&quot;\ny_true\uff1a\u6bcf\u4e2a\u6837\u672c\u7684\u771f\u5b9e\u7c7b\u522b\uff0c\u4ee50\uff08\u53cd\u4f8b\uff09\u30011\uff08\u6b63\u4f8b\uff09\u6807\u6ce8\ny_score\uff1a\u9884\u6d4b\u5f97\u5206\n&quot;&quot;&quot;\n<\/code><\/pre>\n<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n<h2>\u4e03\u3001\u51b3\u7b56\u6811<\/h2>\n<h3>\uff08\u4e00\uff09\u51b3\u7b56\u6811\u7b80\u4ecb<\/h3>\n<h4>1.\u51b3\u7b56\u6811\u7684\u5b9a\u4e49<\/h4>\n<p>\u51b3\u7b56\u6811\uff1a\u4ece\u6839\u7ed3\u70b9\u5f00\u59cb\uff0c\u975e\u53f6\u5b50\u7ed3\u70b9\u4e2d\u5b58\u50a8\u5224\u65ad\u7684\u6761\u4ef6\uff0c\u6839\u636e\u662f\u5426\u6ee1\u8db3\u6761\u4ef6\u8fdb\u5165\u4e0d\u540c\u7684\u5206\u652f\uff0c\u5982\u6b64\u53cd\u590d\uff0c\u76f4\u5230\u53f6\u5b50\u7ed3\u70b9\u3002\u51b3\u7b56\u6811\u662f\u975e\u53c2\u6570\u5b66\u4e60\u7b97\u6cd5\uff0c\u53ef\u4ee5\u89e3\u51b3\u5206\u7c7b\u95ee\u9898\u548c\u56de\u5f52\u95ee\u9898<\/p>\n<h4>2.\u51b3\u7b56\u6811\u7684\u6784\u5efa<\/h4>\n<ul>\n<li>\u7279\u5f81\u9009\u62e9\uff1a\u6311\u9009\u5f3a\u5206\u7c7b\u7684\u7279\u5f81<\/li>\n<li>\u51b3\u7b56\u6811\u751f\u6210<\/li>\n<li>\u51b3\u7b56\u6811\u526a\u679d<\/li>\n<\/ul>\n<h4>3.\u51b3\u7b56\u6811API<\/h4>\n<ul>\n<li><code>from sklearn.tree import DecisionTreeClassifier<\/code>\uff1a\u5206\u7c7b\u5668<\/li>\n<li><code>from skleran.tree import plot_tree<\/code>\uff1a\u7ed8\u56fe<\/li>\n<\/ul>\n<h3>\uff08\u4e8c\uff09ID3\u51b3\u7b56\u6811<\/h3>\n<h4>1.\u4fe1\u606f\u71b5<\/h4>\n<ul>\n<li>\n<p>\u5b9a\u4e49\uff1a\u71b5\u4ee3\u8868\u968f\u673a\u53d8\u91cf\u7684\u4e0d\u786e\u5b9a\u5ea6\u3002\u71b5\u8d8a\u5927\uff0c\u4e0d\u786e\u5b9a\u6027\u8d8a\u9ad8\uff1b\u71b5\u8d8a\u5c0f\uff0c\u4e0d\u786e\u5b9a\u6027\u8d8a\u4f4e<\/p>\n<\/li>\n<li>\n<p>\u516c\u5f0f\uff1a\n$$\nH=-\\sum_{i=1}^k p_i log(p_i)\\\\\n\u5f53\u53ea\u6709\u4e24\u4e2a\u7c7b\u522b\u65f6\uff0c\u7b49\u4ef7\u4e8e\uff1aH=-xlog(x)-(1-x)log(1-x)\n$$<\/p>\n<ul>\n<li>$ p_i  $\uff1a\u8868\u793a\u7b2ci\u4e2a\u7c7b\u522b\u7684\u5728\u603b\u6570\u4e2d\u7684\u5360\u6bd4<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n<h4>2.\u4fe1\u606f\u589e\u76ca<\/h4>\n<ul>\n<li>\n<p>\u5b9a\u4e49\uff1a\u7279\u5f81A\u5bf9\u4e8e\u8bad\u7ec3\u96c6D\u7684\u4fe1\u606f\u589e\u76ca$ g(D,A) $\uff0c\u5b9a\u4e49\u4e3a\u96c6\u5408D\u7684\u7ecf\u9a8c\u71b5$ H(D) $\u4e0e\u7279\u5f81A\u6307\u5b9a\u6761\u4ef6\u4e0bD\u7684\u7ecf\u9a8c\u71b5$ H(D|A) $\u7684\u5dee\n$$\ng(D,A)=H(D)-H(D|A)\n$$<\/p>\n<\/li>\n<li>\n<p>\u7b97\u6cd5\uff1a<\/p>\n<ul>\n<li>\n<p>\u524d\u63d0\u5047\u8bbe\uff1a\n$$\n\\begin{aligned}\n&amp;(1)\u8bbe\u8bad\u7ec3\u6570\u636e\u96c6\u4e3aD\uff0c|D|\u8868\u793a\u5176\u6837\u672c\u7684\u4e2a\u6570.\\\\\n&amp;(2)\u8bbe\u6709K\u4e2a\u7c7bC_k\uff0ck=1,2,\u2026\uff0c|C_k|\u8868\u793a\u5c5e\u4e8e\u7c7bC_k\u7684\u6837\u672c\u4e2a\u6570\uff0c\\sum_{k=1}^K =|D|.\\\\\n&amp;(3)\u8bbe\u7279\u5f81A\u6709n\u4e2a\u4e0d\u540c\u53d6\u503c{a_1,a_2,\u2026,a_n}\uff0c\u6839\u636e\u7279\u5f81A\u7684\u53d6\u503c\u5c06D\u5212\u5206\u4e3an\u4e2a\u5b50\u96c6D_1,D_2,\u2026,D_n\uff0c|D_i|\u8868\u793aD_i\u7684\u6837\u672c\u4e2a\u6570\uff0c\\sum_{i_1}^n|D_i|=|D|.\\\\\n&amp;(4)\u5b50\u96c6\u4e2d\u5c5e\u4e8e\u7c7bC_k\u7684\u6837\u672c\u96c6\u5408\u4e3aD_{ik}\uff0c\u5219\u6709D_{ik}=D_i\\cap C_k\uff0c|D_{ik}|\u4e3aD_{ik}\u7684\u6837\u672c\u4e2a\u6570.\n\\end{aligned}\n$$<\/p>\n<\/li>\n<li>\n<p>\u8f93\u5165\u8f93\u51fa\uff1a<\/p>\n<ul>\n<li>\u8f93\u5165\uff1a\u8bad\u7ec3\u6570\u636e\u96c6D\uff0c\u7279\u5f81A<\/li>\n<li>\u8f93\u51fa\uff1a\u7279\u5f81A\u5bf9\u4e8e\u8bad\u7ec3\u6570\u636e\u96c6D\u7684\u4fe1\u606f\u589e\u76cag(D,A)<\/li>\n<\/ul>\n<\/li>\n<li>\n<p>\u6b65\u9aa4\uff1a<\/p>\n<ul>\n<li>\n<p>\u8ba1\u7b97\u6570\u636e\u96c6D\u7684\u7ecf\u9a8c\u71b5H(D)\uff1a\n$$\nH(D)=-\\sum_{k=1}^K \\frac{|C_k|}{|D|}log\\frac{|C_k|}{|D|}\n$$<\/p>\n<\/li>\n<li>\n<p>\u8ba1\u7b97\u7279\u5f81A\u5bf9\u6570\u636e\u96c6D\u7684\u7ecf\u9a8c\u6761\u4ef6\u71b5H(D|A)\uff1a\n$$\nH(D|A)=\\sum_{i=1}^n \\frac{|D_i|}{|D|}H(D_i)=-\\sum_{i=1}^n \\frac{|D_i|}{|D|}\\sum_{i=1}^n \\frac{|D_{ik}|}{|D_i|}log\\frac{|D_{ik}|}{|D_i|}\n$$<\/p>\n<\/li>\n<li>\n<p>\u8ba1\u7b97\u4fe1\u606f\u589e\u76ca\uff1a\n$$\ng(D|A)=H(D)-H(D|A)\n$$<\/p>\n<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n<\/li>\n<li>\n<p>\u9009\u62e9\uff1a\u9009\u62e9\u4fe1\u606f\u589e\u76ca\u6700\u5927\u7684\u4f5c\u4e3a\u7ed3\u70b9\u7684\u5224\u65ad\u6761\u4ef6\uff0c\u5c06\u6570\u636e\u96c6\u6309\u5224\u65ad\u6761\u4ef6\u5212\u5206\uff0c\u5bf9\u6bcf\u4e00\u4e2a\u5b50\u96c6\u518d\u6b21\u8ba1\u7b97\u4fe1\u606f\u589e\u76ca\uff0c\u5982\u6b64\u5f80\u590d<\/p>\n<\/li>\n<\/ul>\n<h4>3.ID3\u7b97\u6cd5\u6b65\u9aa4<\/h4>\n<ul>\n<li>\u8ba1\u7b97\u6bcf\u4e2a\u7279\u5f81\u7684\u4fe1\u606f\u589e\u76ca<\/li>\n<li>\u4f7f\u7528\u4fe1\u606f\u589e\u76ca\u6700\u5927\u7684\u4f5c\u4e3a\u5224\u65ad\u6761\u4ef6\u62c6\u5206\u6570\u636e\u96c6<\/li>\n<li>\u4f7f\u7528\u5269\u4f59\u7684\u7279\u5f81\u5bf9\u5b50\u96c6\u7ee7\u7eed\u8ba1\u7b97\u4fe1\u606f\u589e\u76ca<\/li>\n<\/ul>\n<h3>\uff08\u4e09\uff09C4.5\u51b3\u7b56\u6811<\/h3>\n<h4>1.\u4fe1\u606f\u589e\u76ca\u7387<\/h4>\n<p>\u516c\u5f0f\uff1a\n$$\nGain_Ratio(D,a)=\\frac{Gain(D,a)}{IV(a)}\\\nIV(a)=-\\sum_{v=1}^V \\frac{D^v}{D}log\\frac{D^v}{D}\n$$<\/p>\n<ul>\n<li>$ Gain_Ratio(D,a) $\uff1a\u8868\u793a\u4fe1\u606f\u589e\u76ca\u7387<\/li>\n<li>$ Gain(D,a) $\uff1a\u8868\u793a\u4fe1\u606f\u589e\u76ca<\/li>\n<li>$ IV(a) $\uff1a\u8868\u793a\u5185\u5728\u4fe1\u606f<\/li>\n<\/ul>\n<h4>2.\u4f7f\u7528\u4fe1\u606f\u589e\u76ca\u7387\u7684\u539f\u56e0<\/h4>\n<p>\u4ec5\u4f7f\u7528\u4fe1\u606f\u589e\u76ca\u4f1a\u4f18\u5148\u9009\u62e9\u5177\u6709\u8f83\u591a\u5c5e\u6027\u503c\u7684\u7279\u5f81\uff0c\u56e0\u4e3a\u5c5e\u6027\u503c\u591a\u7684\u7279\u5f81\u4f1a\u4ea7\u751f\u8f83\u5927\u7684\u4fe1\u606f\u589e\u76ca\uff0c\u4f46\u8fd9\u4e2a\u7279\u5f81\u4e0d\u4e00\u5b9a\u662f\u6700\u4f18\u7684<\/p>\n<h3>\uff08\u56db\uff09CART\u5206\u7c7b\u51b3\u7b56\u6811<\/h3>\n<h4>1.CART\u51b3\u7b56\u6811\u7b80\u4ecb<\/h4>\n<ul>\n<li>CART\u51b3\u7b56\u6811\u662f\u4e00\u79cd\u51b3\u7b56\u6811\u6a21\u578b\uff0c\u53ef\u4ee5\u7528\u4e8e\u5206\u7c7b<code>DecisionTreeClassifier<\/code>\uff0c\u4e5f\u53ef\u4ee5\u7528\u4e8e\u56de\u5f52<code>DecisionTreeRegressor<\/code><\/li>\n<li>CART\u51b3\u7b56\u6811\u7684\u5b66\u4e60\u7b97\u6cd5\u901a\u5e38\u5305\u62ec\u51b3\u7b56\u6811\u751f\u6210\u548c\u51b3\u7b56\u6811\u526a\u679d\uff0c\u7528\u9014\u4e0d\u540c\u7684\u51b3\u7b56\u6811\u4f1a\u91c7\u7528\u4e0d\u540c\u7684\u4f18\u5316\u7b97\u6cd5\n<ul>\n<li>CART\u5206\u7c7b\u6811\uff1a\u4f7f\u7528\u57fa\u5c3c\u6307\u6570\u6700\u5c0f\u5316\u7b56\u7565<\/li>\n<li>CART\u56de\u5f52\u6811\uff1a\u4f7f\u7528\u5e73\u65b9\u8bef\u5dee\u6700\u5c0f\u5316\u7b56\u7565<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n<h4>2.\u57fa\u5c3c\u6307\u6570<\/h4>\n<p>\u516c\u5f0f\uff1a\n$$\nGini(D)=1-\\sum_{k=1}^K p^2\\\\\nGini_index(D,a)=\\sum_{v=1}^V\\frac{D^v}{D}Gini(D^v)\n$$<\/p>\n<ul>\n<li>$ Gini(D) $\uff1a\u57fa\u5c3c\u7cfb\u6570<\/li>\n<li>$ Gini_index(D,a) $\uff1a\u57fa\u5c3c\u6307\u6570<\/li>\n<\/ul>\n<p>\u9009\u62e9\uff1a\u9009\u62e9\u57fa\u5c3c\u6307\u6570\u6700\u5c0f\u7684\u4f5c\u4e3a\u5206\u5272\u70b9<\/p>\n<h4>3.\u751f\u6210CART\u51b3\u7b56\u6811<\/h4>\n<pre><code class=\"language-python\">DecisionTreeClassifier(max_depth=,criterion=)\n\n#max_depth\uff1a\u8868\u793a\u51b3\u7b56\u6811\u6700\u5927\u6df1\u5ea6\n#criterion\uff1a\u8868\u793a\u51b3\u7b56\u6811\u7c7b\u578b\uff0c\u9ed8\u8ba4\u4e3a&quot;gini&quot;\u751f\u6210CART\u51b3\u7b56\u6811\n<\/code><\/pre>\n<h3>\uff08\u4e94\uff09CART\u56de\u5f52\u51b3\u7b56\u6811<\/h3>\n<h4>1.\u533a\u522b<\/h4>\n<table>\n<thead>\n<tr>\n<th style=\"text-align:center\">CART\u5206\u7c7b\u51b3\u7b56\u6811<\/th>\n<th style=\"text-align:center\">CART\u56de\u5f52\u51b3\u7b56\u6811<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td style=\"text-align:center\">\u9884\u6d4b\u8f93\u51fa\u4e00\u4e2a\u79bb\u6563\u503c<\/td>\n<td style=\"text-align:center\">\u9884\u6d4b\u8f93\u51fa\u4e00\u4e2a\u8fde\u7eed\u503c<\/td>\n<\/tr>\n<tr>\n<td style=\"text-align:center\">\u4f7f\u7528\u57fa\u5c3c\u6307\u6570\u6784\u5efa<\/td>\n<td style=\"text-align:center\">\u4f7f\u7528\u5e73\u65b9\u635f\u5931\u6784\u5efa<\/td>\n<\/tr>\n<tr>\n<td style=\"text-align:center\">\u4f7f\u7528\u53f6\u5b50\u8282\u70b9\u51fa\u73b0\u66f4\u591a\u6b21\u6570\u7684\u7c7b\u522b\u4f5c\u4e3a\u9884\u6d4b\u7c7b\u522b<\/td>\n<td style=\"text-align:center\">\u4f7f\u7528\u53f6\u5b50\u8282\u70b9\u91cc\u7684\u5747\u503c\u4f5c\u4e3a\u9884\u6d4b\u8f93\u51fa<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<h4>2.\u5e73\u65b9\u635f\u5931<\/h4>\n<p>\u516c\u5f0f\uff1a\n$$\n\\begin{align}\n&amp;\u8bbe\u5b58\u5728n\u4e2a\u7279\u5f81\uff0c\u5176\u4e2d\u7b2ci\u4e2a\u7279\u5f81\u7684\u76ee\u6807\u503c\u4e3ay_i.\u5bf9\u7279\u5f81\u8fdb\u884c\u6392\u5e8f.\\\\\n&amp;\u8ba1\u7b97\u6bcf\u76f8\u90bb2\u4e2a\u7279\u5f81\u7684\u5747\u503c\u4f5c\u4e3a\u5212\u5206\u70b9\uff0c\u4e00\u5171\u53ef\u4ee5\u5f62\u6210n-1\u4e2a\u5212\u5206\u70b9.\\\\\n&amp;\u5bf9\u4e8e\u6bcf\u4e00\u4e2a\u5212\u5206\u70b9\uff0c\u5de6\u4fa7\u7684\u7279\u5f81\u8bb0\u4e3a\u96c6\u5408L\uff0c\u53f3\u4fa7\u7684\u7279\u5f81\u8bb0\u4e3a\u96c6\u5408R.L_i\u8868\u793aL\u4e2d\u7b2ci\u4e2a\u7279\u5f81\u5bf9\u5e94\u7684\u76ee\u6807\u503c\uff0cR_i\u540c\u7406.\\\\\n&amp;\u5bf9L\u548cR\u4e2d\u7684\u76ee\u6807\u503c\u6c42\u5747\u503c\uff0c\u5f97\u5230\\overline L\u548c\\overline R.\\\\\n&amp;\u8be5\u5212\u5206\u70b9\u7684\u5e73\u65b9\u635f\u5931:Loss=\\sum_{i=1}^L (L_i-\\overline L)^2+\\sum_{j=1}^{R}(R_j-\\overline R)^2\n\\end{align}\n$$\n\u9009\u62e9\uff1a\u5e73\u65b9\u635f\u5931\u6700\u5c0f\u7684\u4f5c\u4e3a\u6700\u7ec8\u7684\u5212\u5206\u70b9<\/p>\n<h3>\uff08\u516d\uff09\u526a\u679d<\/h3>\n<h4>1.\u9884\u526a\u679d<\/h4>\n<ul>\n<li>\u9884\u526a\u679d\uff1a\u662f\u5728\u51b3\u7b56\u6811\u751f\u6210\u5b8c\u6210\u4e4b\u524d\u8fdb\u884c\u7684\u526a\u679d\u64cd\u4f5c\uff0c\u9650\u5236\u51b3\u7b56\u6811\u7684\u751f\u6210\u89c4\u6a21<\/li>\n<li>\u65b9\u5f0f\uff1a\u5224\u65ad\u4e00\u4e2a\u8282\u70b9\u5728\u5206\u88c2\u4e4b\u540e\u80fd\u5426\u63d0\u5347\u6a21\u578b\u7684\u51c6\u786e\u7387\uff0c\u82e5\u4e0d\u80fd\u5219\u526a\u6389<\/li>\n<li>\u4f18\u70b9\uff1a\u4f7f\u51b3\u7b56\u6811\u5f88\u591a\u5206\u652f\u4e0d\u5c55\u5f00\uff0c\u964d\u4f4e\u4e86\u8fc7\u62df\u5408\u98ce\u9669\u548c\u8bad\u7ec3\u5f00\u9500<\/li>\n<li>\u7f3a\u70b9\uff1a\n<ul>\n<li>\u6709\u4e9b\u5206\u652f\u867d\u7136\u4e0d\u80fd\u63d0\u5347\u6a21\u578b\u51c6\u786e\u7387\uff0c\u4f46\u53ef\u80fd\u5176\u540e\u7eed\u5206\u652f\u80fd\u63d0\u5347\u51c6\u786e\u7387<\/li>\n<li>\u5e26\u6765\u4e86\u6b20\u62df\u5408\u98ce\u9669<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n<h4>2.\u540e\u526a\u679d<\/h4>\n<ul>\n<li>\u540e\u526a\u679d\uff1a\u662f\u5728\u51b3\u7b56\u6811\u751f\u6210\u5b8c\u6210\u4e4b\u540e\u8fdb\u884c\u7684\u526a\u679d\u64cd\u4f5c\uff0c\u524a\u51cf\u51b3\u7b56\u6811\u7684\u89c4\u6a21<\/li>\n<li>\u65b9\u5f0f\uff1a\u5224\u65ad\u4e00\u4e2a\u8282\u70b9\u5728\u88ab\u526a\u9664\u4e4b\u540e\u80fd\u5426\u63d0\u5347\u6a21\u578b\u7684\u51c6\u786e\u7387\uff0c\u82e5\u80fd\u5219\u526a\u6389<\/li>\n<li>\u4f18\u70b9\uff1a\u4fdd\u7559\u4e86\u66f4\u591a\u5206\u652f\uff0c\u4e00\u822c\u60c5\u51b5\u4e0b\u540e\u526a\u679d\u5bfc\u81f4\u6b20\u62df\u5408\u7684\u98ce\u9669\u8f83\u5c0f\u4e00\u4e9b<\/li>\n<li>\u7f3a\u70b9\uff1a\u8bad\u7ec3\u5f00\u9500\u5927<\/li>\n<\/ul>\n<h2>\u516b\u3001\u6734\u7d20\u8d1d\u53f6\u65af<\/h2>\n<h3>\uff08\u4e00\uff09\u5e38\u7528\u6982\u7387<\/h3>\n<p>\u6761\u4ef6\u6982\u7387\uff1a\u8868\u793a\u4e8b\u4ef6A\u5728\u4e8b\u4ef6B\u5df2\u7ecf\u53d1\u751f\u7684\u6761\u4ef6\u4e0b\u53d1\u751f\u7684\u6982\u7387\uff0c$ P(A|B) $<\/p>\n<p>\u8054\u5408\u6982\u7387\uff1a\u8868\u793a\u591a\u4e2a\u6761\u4ef6\u540c\u65f6\u6210\u7acb\u7684\u6982\u7387\uff0c$ P(AB)=P(A)P(B|A) $<\/p>\n<h3>\uff08\u4e8c\uff09\u8d1d\u53f6\u65af\u516c\u5f0f<\/h3>\n<p>$$\nP(A|B)=\\frac{P(B|A)P(A)}{P(B)}\\\n\u7b49\u4ef7\u4e8e:P(A|B)P(B)=P(B|A)P(A)\n$$<\/p>\n<h3>\uff08\u4e09\uff09\u62c9\u666e\u62c9\u65af\u5e73\u6ed1\u7cfb\u6570<\/h3>\n<p>\u7531\u4e8e\u8bad\u7ec3\u6837\u672c\u53ef\u80fd\u4e0d\u8db3\uff0c\u5bfc\u81f4\u6982\u7387\u8ba1\u7b97\u4e2d\u51fa\u73b00\u7684\u60c5\u51b5\uff0c\u4e3a\u4e86\u89e3\u51b3\u8fd9\u4e00\u95ee\u9898\uff0c\u6211\u4eec\u5f15\u5165\u62c9\u666e\u62c9\u65af\u5e73\u6ed1\u7cfb\u6570\uff1a\n$$\nP(A|B)=\\frac{N_{AB}+\\alpha}{N_{B}+\\alpha m}\n$$<\/p>\n<ul>\n<li>$ N_{AB} $\uff1a\u8868\u793aA\u4e2d\u7b26\u5408\u6761\u4ef6B\u7684\u6837\u672c\u7684\u4e2a\u6570<\/li>\n<li>$ N_B $\uff1a\u8868\u793a\u6761\u4ef6B\u4e0b\u6240\u6709\u6837\u672c\u7684\u4e2a\u6570<\/li>\n<li>$ m $\uff1a\u8868\u793a\u6240\u6709\u72ec\u7acb\u6837\u672c\u7684\u603b\u6570<\/li>\n<li>$ \\alpha $\uff1a\u8868\u793a\u62c9\u666e\u62c9\u65af\u5e73\u6ed1\u7cfb\u6570\uff0c\u4e00\u822c\u6307\u5b9a\u4e3a1<\/li>\n<\/ul>\n<h2>\u4e5d\u3001\u652f\u6301\u5411\u91cf\u673a<\/h2>\n<h3>\uff08\u4e00\uff09\u652f\u6301\u5411\u91cf<\/h3>\n<h4>1.\u6982\u5ff5<\/h4>\n<ul>\n<li>\u652f\u6301\u5411\u91cf\u673a\u7684\u57fa\u672c\u6a21\u578b\u662f\u5b9a\u4e49\u5728\u7279\u5f81\u7a7a\u95f4\u4e0a\u7684\u95f4\u9694\u6700\u5927\u7684\u7ebf\u6027\u5206\u7c7b\u5668\uff0c\u662f\u4e00\u79cd\u4e8c\u5206\u7c7b\u7684\u6a21\u578b\uff0c\u5f53\u4f7f\u7528\u4e86\u6838\u6280\u5de7\u4e4b\u540e\uff0c\u652f\u6301\u5411\u91cf\u673a\u53ef\u4ee5\u7528\u4e8e\u975e\u7ebf\u6027\u5206\u7c7b<\/li>\n<li>SVM\uff1a\u6307\u7684\u662fN\u7ef4\u7a7a\u95f4\u7684\u5206\u7c7b\u8d85\u5e73\u9762\uff0c\u5176\u5c06\u7a7a\u95f4\u5206\u4e3a\u4e24\u90e8\u5206<\/li>\n<li>\u652f\u6301\u5411\u91cf\uff1a\u6307\u7684\u662f\u843d\u5728\u8fb9\u9645\u4e24\u8fb9\u7684\u8d85\u5e73\u9762\u4e0a\u7684\u70b9\uff0c\u7528\u4e8e\u652f\u6301\u6784\u5efa\u6700\u5927\u8fb9\u7f18\u8d85\u5e73\u9762<\/li>\n<li><code>from sklearn.svm import SVC<\/code><\/li>\n<\/ul>\n<h4>2.\u5206\u7c7b<\/h4>\n<ul>\n<li>\u7ebf\u6027\u53ef\u5206\u652f\u6301\u5411\u91cf\u673a\uff08\u786c\u95f4\u9694\u652f\u6301\u5411\u91cf\u673a\uff09\n<ul>\n<li>\u5f53\u8bad\u7ec3\u6570\u636e\u7ebf\u6027\u53ef\u5206\u65f6\uff0c\u901a\u8fc7\u786c\u95f4\u9694\u6700\u5927\u5316\uff0c\u5b66\u4e60\u5f97\u5230\u4e00\u4e2a\u7ebf\u6027\u53ef\u5206\u652f\u6301\u5411\u91cf\u673a<\/li>\n<\/ul>\n<\/li>\n<li>\u7ebf\u6027\u652f\u6301\u5411\u91cf\u673a\uff08\u8f6f\u95f4\u9694\u652f\u6301\u5411\u91cf\u673a\uff09\n<ul>\n<li>\u5f53\u8bad\u7ec3\u6570\u636e\u8fd1\u4f3c\u7ebf\u6027\u53ef\u5206\u65f6\uff0c\u901a\u8fc7\u8f6f\u95f4\u9694\u6700\u5927\u5316\uff0c\u5b66\u4e60\u5f97\u5230\u4e00\u4e2a\u7ebf\u6027\u652f\u6301\u5411\u91cf\u673a<\/li>\n<\/ul>\n<\/li>\n<li>\u975e\u7ebf\u6027\u652f\u6301\u5411\u91cf\u673a\n<ul>\n<li>\u5f53\u8bad\u7ec3\u6570\u636e\u4e0d\u53ef\u5206\u65f6\uff0c\u901a\u8fc7\u4f7f\u7528\u6838\u6280\u5de7\u4ee5\u53ca\u8f6f\u95f4\u9694\u6700\u5927\u5316\uff0c\u5b66\u4e60\u5f97\u5230\u4e00\u4e2a\u975e\u7ebf\u6027\u652f\u6301\u5411\u91cf\u673a<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n<h3>\uff08\u4e8c\uff09\u8f6f\u95f4\u9694\u548c\u786c\u95f4\u9694<\/h3>\n<ul>\n<li>\u786c\u95f4\u9694\uff1a\u6307\u7684\u662f\u8ba9\u6240\u6709\u6837\u672c\u90fd\u4e0d\u5728\u6700\u5927\u95f4\u9694\u4e4b\u95f4\uff0c\u5e76\u5904\u4e8e\u6b63\u786e\u7684\u4e00\u4fa7\u3002\u4f46\u5982\u679c\u51fa\u73b0\u5f02\u5e38\u503c\u6216\u6837\u672c\u4e0d\u80fd\u7ebf\u6027\u53ef\u5206\uff0c\u5219\u786c\u95f4\u9694\u65e0\u6cd5\u5b9e\u73b0<\/li>\n<li>\u8f6f\u95f4\u9694\uff1a\u6307\u7684\u662f\u53ef\u4ee5\u5bb9\u5fcd\u4e00\u90e8\u5206\u6837\u672c\u5728\u6700\u5927\u95f4\u9694\u4e4b\u5185\uff0c\u751a\u81f3\u5728\u9519\u8bef\u7684\u4e00\u4fa7<\/li>\n<\/ul>\n<h3>\uff08\u4e09\uff09\u60e9\u7f5a\u53c2\u6570C<\/h3>\n<p>\u5728\u8f6f\u95f4\u9694\u60c5\u51b5\u4e0b\uff0c\u6211\u4eec\u9700\u8981\u8003\u8651\u6700\u5927\u5316\u95f4\u9694\u548c\u90a3\u4e9b\u9650\u5236\u95f4\u9694\u8fdd\u4f8b\u7684\u6837\u672c\u6240\u5e26\u6765\u7684\u635f\u5931<\/p>\n<ul>\n<li>C\u8d8a\u5927\uff1a\u8868\u793a\u8fdd\u53cd\u9650\u5236\u95f4\u9694\u7684\u6837\u672c\u70b9\u5e26\u6765\u7684\u635f\u5931\u8d8a\u5927\uff0c\u8981\u51cf\u5c11\u6b64\u7c7b\u6837\u672c\u7684\u6570\u91cf\uff0c\u4e5f\u5c31\u662f\u51cf\u5c0f\u95f4\u9694<\/li>\n<li>C\u8d8a\u5c0f\uff1a\u8868\u793a\u8fdd\u53cd\u9650\u5236\u95f4\u9694\u7684\u6837\u672c\u70b9\u5e26\u6765\u7684\u635f\u5931\u8d8a\u5c0f\uff0c\u53ef\u4ee5\u9002\u5f53\u589e\u5927\u95f4\u9694\u6765\u63d0\u9ad8\u6a21\u578b\u7684\u6cdb\u5316\u80fd\u529b<\/li>\n<\/ul>\n<h3>\uff08\u56db\uff09SVM\u6838\u65b9\u6cd5\u548c\u635f\u5931\u51fd\u6570<\/h3>\n<h4>1.\u6838\u65b9\u6cd5<\/h4>\n<p>\u5f53\u5b58\u5728\u7ebf\u6027\u4e0d\u53ef\u5206\u7684\u60c5\u51b5\u65f6\uff0c\u6211\u4eec\u9700\u8981\u4f7f\u7528\u6838\u51fd\u6570\u63d0\u9ad8\u8bad\u7ec3\u6837\u672c\u7684\u7ef4\u5ea6\u6216\u5c06\u8bad\u7ec3\u6837\u672c\u6295\u5411\u9ad8\u7ef4\n$$\n\\begin{align}\n&amp;\u7ebf\u6027\u6838:k(x_i,x_j)=x_i^Tx_j\\\\\n&amp;\u591a\u9879\u5f0f\u6838:k(x_i,x_j)=(x_i^Tx_j)^d,[d&gt;=1\u4e3a\u591a\u9879\u5f0f\u7684\u6b21\u6570]\\\\\n&amp;\u9ad8\u65af\u6838:k(x_i,x_j)=exp(-\\frac{||x_i-x_j||^2}{2\\sigma^2}),[\\sigma&gt;0\u4e3a\u9ad8\u65af\u6838\u7684\u5e26\u5bbd]\\\\\n&amp;\u62c9\u666e\u62c9\u65af\u6838:k(x_i,x_j)=exp(-\\frac{||x_i-x_j||}{\\sigma}),[\\sigma&gt;0]\\\\\n&amp;Sigmoid\u6838:k(x_i,x_j)=tanh(\\beta x_i^Tx_j+\\theta),[tanh\u4e3a\u53cc\u66f2\u6b63\u5207\u51fd\u6570,\\beta&gt;0,\\theta&lt;0]\n\\end{align}\n$$<\/p>\n<ul>\n<li>\u7ebf\u6027\u6838\uff1a\u4e00\u822c\u4e0d\u589e\u52a0\u6570\u636e\u7684\u7ef4\u5ea6\uff0c\u800c\u662f\u9884\u5148\u8ba1\u7b97\u5185\u79ef<\/li>\n<li>\u591a\u9879\u5f0f\u6838\uff1a\u4e00\u822c\u662f\u901a\u8fc7\u589e\u52a0\u591a\u9879\u5f0f\u7279\u5f81\uff0c\u63d0\u5347\u6570\u636e\u7ef4\u5ea6\u5e76\u8ba1\u7b97\u5185\u79ef<\/li>\n<li>\u9ad8\u65af\u6838\uff1a\u4e00\u822c\u662f\u5c06\u6837\u672c\u6295\u5c04\u5230\u65e0\u9650\u7ef4\u7a7a\u95f4\uff0c\u4f7f\u5f97\u539f\u672c\u4e0d\u53ef\u5206\u7684\u6570\u636e\u53d8\u5f97\u53ef\u5206<\/li>\n<\/ul>\n<h4>2.\u635f\u5931\u51fd\u6570<\/h4>\n<ul>\n<li>0-1\u635f\u5931\uff1a\n<ul>\n<li>\u5f53\u6b63\u4f8b\u6837\u672c\u843d\u5728y=0\u4e0b\u65b9\u65f6\u635f\u5931\u4e3a0\uff0c\u5426\u5219\u4e3a1<\/li>\n<li>\u5f53\u8d1f\u4f8b\u6837\u672c\u843d\u5728y=0\u4e0a\u65b9\u65f6\u635f\u5931\u4e3a0\uff0c\u5426\u5219\u4e3a1<\/li>\n<\/ul>\n<\/li>\n<li>Hinge\u635f\u5931\uff1a\n<ul>\n<li>\u5f53\u6b63\u4f8b\u6837\u672c\u843d\u5728y&gt;=1\u4e00\u4fa7\u65f6\u635f\u5931\u4e3a0\uff0c\u5426\u5219\u8ddd\u79bb\u8d8a\u8fdc\u635f\u5931\u8d8a\u5927<\/li>\n<li>\u5f53\u8d1f\u4f8b\u6837\u672c\u843d\u5728y&lt;=-1\u4e00\u4fa7\u65f6\u635f\u5931\u4e3a0\uff0c\u5426\u5219\u8ddd\u79bb\u8d8a\u8fdc\u635f\u5931\u8d8a\u5927<\/li>\n<\/ul>\n<\/li>\n<li>Logistic\u635f\u5931\uff1a\n<ul>\n<li>\u5f53\u6b63\u4f8b\u6837\u672c\u843d\u5728y&gt;0\u4e00\u4fa7\u65f6\uff0c\u8ddd\u79bby=0\u8d8a\u8fdc\u5219\u635f\u5931\u8d8a\u5c0f<\/li>\n<li>\u5f53\u8d1f\u4f8b\u6837\u672c\u843d\u5728y&lt;0\u4e00\u4fa7\u65f6\uff0c\u8ddd\u79bby=0\u8d8a\u8fdc\u5219\u635f\u5931\u8d8a\u5c0f<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n<h2>\u5341\u3001\u805a\u7c7b\u7b97\u6cd5<\/h2>\n<h3>\uff08\u4e00\uff09\u6982\u5ff5<\/h3>\n<p>\u805a\u7c7b\u7b97\u6cd5\uff1a\u662f\u4e00\u79cd\u5178\u578b\u7684\u65e0\u76d1\u7763\u5b66\u4e60\u7b97\u6cd5\uff0c\u4e3b\u8981\u7528\u4e8e\u5c06\u76f8\u4f3c\u7684\u6837\u672c\u5212\u5206\u5230\u540c\u4e00\u7c7b\u522b\u4e2d<\/p>\n<h3>\uff08\u4e8c\uff09\u805a\u7c7b\u7b97\u6cd5\u7684API<\/h3>\n<h4>1.\u5bfc\u5305<\/h4>\n<p><code>sklearn.cluster.KMeans<\/code><\/p>\n<h4>2.\u53c2\u6570<\/h4>\n<p><code>n_clusters=<\/code>\uff1a\u6307\u5b9a\u805a\u7c7b\u7684\u6570\u91cf<\/p>\n<p><code>init=<\/code>\uff1a\u53ef\u4ee5\u586b\u5165\u2019k-means++&#8217;\uff0c\u8868\u793a\u9009\u62e9\u4e0b\u4e00\u4e2a\u4e2d\u5fc3\u70b9\u65f6\uff0c\u79bb\u73b0\u6709\u7684\u4e2d\u5fc3\u70b9\u8fdc\u7684\u70b9\u7684\u9009\u62e9\u6982\u7387\u8f83\u9ad8<\/p>\n<h3>\uff08\u4e09\uff09k-means\u805a\u7c7b\u6b65\u9aa4<\/h3>\n<ul>\n<li>\u968f\u673a\u8bbe\u7f6ek\u4e2a\u7279\u5f81\u7a7a\u95f4\u5185\u7684\u70b9\u4f5c\u4e3a\u521d\u59cb\u7684\u805a\u7c7b\u4e2d\u5fc3<\/li>\n<li>\u5bf9\u4e8e\u5176\u4ed6\u7684\u70b9\u8ba1\u7b97\u5230k\u4e2a\u4e2d\u5fc3\u7684\u8ddd\u79bb\uff0c\u672a\u77e5\u7684\u70b9\u9009\u62e9\u6700\u8fd1\u7684\u4e2d\u5fc3\u70b9\u4f5c\u4e3a\u6807\u8bb0\u7c7b\u522b<\/li>\n<li>\u5bf9\u4e8e\u6bcf\u4e00\u4e2a\u805a\u7c7b\uff0c\u8ba1\u7b97\u5e73\u5747\u503c\u4f5c\u4e3a\u65b0\u7684\u4e2d\u5fc3\u70b9<\/li>\n<li>\u5982\u679c\u65b0\u4e2d\u5fc3\u70b9\u4e0e\u539f\u4e2d\u5fc3\u70b9\u4e00\u81f4\uff0c\u5219\u7ed3\u675f\uff1b\u5426\u5219\u91cd\u590d\u7b2c\u4e8c\u6b65<\/li>\n<\/ul>\n<h3>\uff08\u56db\uff09\u6a21\u578b\u8bc4\u4f30<\/h3>\n<h4>1.\u8bef\u5dee\u5e73\u65b9\u548c\uff08SSE\uff09<\/h4>\n<p>$$\nSSE=\\sum_{i=1}^k \\sum_{p\\in C_i} |p-m_i|^2\n$$<\/p>\n<ul>\n<li>\u7b80\u5355\u6765\u8bf4\uff0c\u5c31\u662f\u5bf9\u4e8e\u6bcf\u4e00\u4e2a\u805a\u7c7b\uff0c\u5c06\u6240\u6709\u6837\u672c\u4e0e\u4e2d\u5fc3\u70b9\u6c42\u8ddd\u79bb\uff0c\u5206\u522b\u5e73\u65b9\u540e\u6c42\u548c<\/li>\n<li>\u6700\u540e\u5c06\u6240\u6709\u805a\u7c7b\u7684\u7ed3\u679c\u52a0\u5230\u4e00\u8d77<\/li>\n<\/ul>\n<h4>2.\u201c\u8098\u201d\u65b9\u6cd5<\/h4>\n<ul>\n<li>\u5bf9\u4e8en\u4e2a\u70b9\u7684\u6570\u636e\u96c6\uff0c\u8fed\u4ee3\u8ba1\u7b97k\uff0ck\u4ece1\u5230n.\u6bcf\u6b21\u805a\u7c7b\u5b8c\u6210\u540e\u8ba1\u7b97\u6bcf\u4e00\u4e2a\u70b9\u5230\u5176\u6240\u5c5e\u4e2d\u5fc3\u7684\u8ddd\u79bb\u5e73\u65b9\u548c<\/li>\n<li>\u5f53k\u9010\u6e10\u589e\u5927\u65f6\uff0c\u8ddd\u79bb\u5e73\u65b9\u548c\u4f1a\u9010\u6e10\u7f29\u5c0f\uff0c\u76f4\u5230k=n\u65f6\uff0c\u8ddd\u79bb\u5e73\u65b9\u548c\u4e3a0<\/li>\n<li>\u5728\u5e73\u65b9\u548c\u53d8\u5316\u8fc7\u7a0b\u4e2d\uff0c\u4f1a\u51fa\u73b0\u4e00\u4e2a\u201c\u62d0\u70b9\u201d\uff0c\u6b64\u70b9\u524d\u540e\u5e73\u65b9\u548c\u7684\u4e0b\u964d\u901f\u5ea6\u53d8\u5316\u8f83\u5927\uff0c\u4e5f\u5c31\u662f\u7a81\u7136\u4e0b\u964d\u901f\u5ea6\u53d8\u7f13\u6162\uff0c\u6211\u4eec\u8ba4\u4e3a\u8be5\u70b9\u662f\u6700\u4f73k\u503c<\/li>\n<\/ul>\n<h4>3.\u8f6e\u5ed3\u7cfb\u6570\u6cd5<\/h4>\n<p>$$\nS=\\frac{(b-a)}{max(a,b)}\n$$<\/p>\n<ul>\n<li>$ a $\uff1a\u8868\u793a\u6837\u672ci\u5230\u540c\u4e00\u7c07\u5185\u5176\u4ed6\u70b9\u4e0d\u76f8\u4f3c\u7a0b\u5ea6\u7684\u5e73\u5747\u503c\uff0c\u8bf4\u767d\u4e86\u5c31\u662f\u4e0e\u81ea\u5df1\u540c\u4e00\u7c07\u7684\u5176\u4ed6\u70b9\u6c42\u8ddd\u79bb\u7684\u5e73\u5747\u503c<\/li>\n<li>$ b $\uff1a\u8868\u793a\u6837\u672ci\u5230\u5176\u4ed6\u7c07\u5185\u5e73\u5747\u4e0d\u76f8\u4f3c\u7a0b\u5ea6\u7684\u6700\u5c0f\u503c\uff0c\u8bf4\u767d\u4e86\u5c31\u662f\u4e0e\u5176\u4ed6\u7c07\u7684\u70b9\u6309\u7167\u7c07\u6c42\u8ddd\u79bb\u7684\u5e73\u5747\u503c\uff0c\u7136\u540e\u53d6\u6700\u5c0f\u503c<\/li>\n<li>$ S $\uff1a\u53d6\u503c\u4e3a[-1,1]\uff0c\u8d8a\u63a5\u8fd11\uff0c\u8868\u793a\u805a\u5408\u7a0b\u5ea6\u8d8a\u597d\uff1b\u8d8a\u63a5\u8fd10\uff0c\u8868\u793a\u8d8a\u9760\u8fd1\u5206\u7c07\u7684\u8fb9\u754c\uff1b\u4e3a\u8d1f\u503c\u65f6\uff0c\u8868\u793a\u8be5\u70b9\u53ef\u80fd\u88ab\u8bef\u5206\u4e86<\/li>\n<\/ul>\n<h3>\uff08\u4e94\uff09\u7279\u5f81\u964d\u7ef4<\/h3>\n<h4>1.\u6982\u5ff5<\/h4>\n<ul>\n<li>\u7528\u4e8e\u8bad\u7ec3\u7684\u6570\u636e\u96c6\u7684\u8d28\u91cf\u5bf9\u6a21\u578b\u7684\u6027\u80fd\u53ca\u5176\u91cd\u8981\u3002\u5982\u679c\u8bad\u7ec3\u96c6\u4e2d\u5305\u62ec\u4e00\u4e9b\u4e0d\u91cd\u8981\u7684\u7279\u5f81\uff0c\u53ef\u80fd\u5bfc\u81f4\u6a21\u578b\u7684\u6cdb\u5316\u6027\u80fd\u4e0d\u4f73<\/li>\n<li>\u964d\u7ef4\u662f\u6307\u5728\u67d0\u4e9b\u9650\u5b9a\u6761\u4ef6\u4e0b\uff0c\u964d\u4f4e\u7279\u5f81\u7684\u4e2a\u6570<\/li>\n<\/ul>\n<h4>2.\u4f4e\u65b9\u5dee\u8fc7\u6ee4<\/h4>\n<ul>\n<li>\n<p>\u82e5\u4e00\u4e2a\u7279\u5f81\u7684\u65b9\u5dee\u5f88\u5c0f\uff0c\u8bf4\u660e\u8fd9\u4e2a\u7279\u5f81\u5305\u542b\u7684\u4fe1\u606f\u8f83\u5c11\uff0c\u6211\u4eec\u8bbe\u5b9a\u4e00\u4e2a\u9608\u503c\uff0c\u8fc7\u6ee4\u6389\u65b9\u5dee\u4f4e\u4e8e\u9608\u503c\u7684\u7279\u5f81<\/p>\n<pre><code class=\"language-python\">\u8fc7\u6ee4\u5bf9\u8c61 = sklearn.feature_selection.VarianceThreshold(threshold=0.0)\n\u6570\u636e = \u8fc7\u6ee4\u5bf9\u8c61.fit_transform(\u6570\u636e)\n\n#threshold\uff1a\u9608\u503c\uff0c\u9ed8\u8ba4\u4e3a0\n<\/code><\/pre>\n<\/li>\n<\/ul>\n<h4>3.\u76f8\u5173\u7cfb\u6570\u6cd5<\/h4>\n<ul>\n<li>\n<p>\u76ae\u5c14\u900a\u76f8\u5173\u7cfb\u6570\uff1a\n$$\nr=\\frac{n\\sum xy-\\sum x \\sum y}{\\sqrt{n\\sum x^2-(\\sum x)^2}\\sqrt{n\\sum y^2-(\\sum y)^2}}\n$$<\/p>\n<\/li>\n<li>\n<p>\u65af\u76ae\u5c14\u66fc\u76f8\u5173\u7cfb\u6570\uff1a\n$$\nRankIC=1-\\frac{6\\sum d_i^2}{n(n^2-1)}\n$$<\/p>\n<\/li>\n<li>\n<p>$ n $\uff1a\u8868\u793a\u7ef4\u5ea6\u6570<\/p>\n<\/li>\n<li>\n<p>$ d_i $\uff1a\u8868\u793a\u6837\u672c\u4e2d\u4e0d\u540c\u7279\u5f81\u5728\u6570\u636e\u4e2d\u6392\u5e8f\u7684\u5e8f\u53f7\u5dee\u503c<\/p>\n<pre><code class=\"language-python\">#\u76ae\u5c14\u900a\u76f8\u5173\u7cfb\u6570\nfrom scipy.stats import pearsonr \npearsonr(data[],data[])\n\n#\u65af\u76ae\u5c14\u66fc\u76f8\u5173\u7cfb\u6570\nfrom scipy.stats import spearmanr\nspearmanr(data[],data[])\n\n#\u76f8\u5173\u7cfb\u6570\ndata.corr(method=&quot;spearmanr&quot;)\n#\u6bcf\u4e00\u4e2a\u7279\u5f81\u4e24\u4e24\u8ba1\u7b97\u76f8\u5173\u7cfb\u6570\uff0c\u9ed8\u8ba4\u4f7f\u7528spearmanr\n<\/code><\/pre>\n<\/li>\n<\/ul>\n<h4>4.\u4e3b\u6210\u5206\u5206\u6790\uff08PCA\uff09<\/h4>\n<p>PCA\u901a\u8fc7\u5bf9\u6570\u636e\u7ef4\u6570\u8fdb\u884c\u538b\u7f29\uff0c\u5c3d\u53ef\u80fd\u964d\u4f4e\u539f\u6570\u636e\u7684\u7ef4\u6570\uff0c\u635f\u5931\u5c11\u91cf\u4fe1\u606f\uff0c\u5728\u6b64\u8fc7\u7a0b\u4e2d\u53ef\u80fd\u4f1a\u820d\u5f03\u539f\u6709\u6570\u636e\u3001\u521b\u9020\u65b0\u7684\u53d8\u91cf<\/p>\n<pre><code class=\"language-python\">from sklearn.decomposition import PCA\n\n\u5bf9\u8c61 = PCA(n_components=)\n\u6570\u636e = \u5bf9\u8c61.fit_transform(\u6570\u636e)\n\n#n_components\uff1a\n#1.\u8868\u793a\u4fdd\u7559\u767e\u5206\u4e4b\u591a\u5c11\u7684\u539f\u59cb\u4fe1\u606f\uff0c\u59820.95\u8868\u793a\u4fdd\u755995\n#2.\u82e5\u4f20\u5165\u7684\u6570\u636e\u5927\u4e8e1\uff0c\u8868\u793a\u4fdd\u7559\u591a\u5c11\u5217\u7684\u4fe1\u606f\uff0c\u59823\u8868\u793a\u4fdd\u75593\u5217\n<\/code><\/pre>\n<h2>\u5341\u4e00\u3001\u96c6\u6210\u5b66\u4e60<\/h2>\n<h3>\uff08\u4e00\uff09\u96c6\u6210\u5b66\u4e60<\/h3>\n<h4>1.\u6982\u5ff5<\/h4>\n<p>\u96c6\u6210\u5b66\u4e60\u7b97\u6cd5\u662f\u5c06\u591a\u4e2a\u5206\u7c7b\u5668\u7ec4\u5408\uff0c\u4ece\u800c\u5b9e\u73b0\u4e00\u4e2a\u9884\u6d4b\u6548\u679c\u66f4\u597d\u7684\u96c6\u6210\u5206\u7c7b\u5668<\/p>\n<h4>2.\u5206\u7c7b<\/h4>\n<ul>\n<li>\u5e76\u884c\u7b97\u6cd5\uff1a\u5229\u7528\u76f8\u540c\u7684\u6570\u636e\u96c6\u540c\u65f6\u642d\u5efa\u591a\u4e2a\u6a21\u578b\uff0c\u7136\u540e\u901a\u8fc7\u6295\u7968\u7684\u65b9\u5f0f\uff0c\u5c11\u6570\u670d\u4ece\u591a\u6570\u5f97\u5230\u6700\u7ec8\u7684\u9884\u6d4b\u7ed3\u679c<\/li>\n<li>\u4e32\u884c\u7b97\u6cd5\uff1a\u6309\u4e00\u5b9a\u6b21\u5e8f\u642d\u5efa\u591a\u4e2a\u6a21\u578b\uff0c\u540e\u52a0\u5165\u7684\u6a21\u578b\u9700\u8981\u5bf9\u73b0\u6709\u7684\u6a21\u578b\u6709\u4e00\u5b9a\u7684\u8d21\u732e\uff0c\u4ece\u800c\u4e0d\u65ad\u63d0\u5347\u6027\u80fd<\/li>\n<\/ul>\n<h4>3.\u6027\u80fd\u8bc4\u4f30<\/h4>\n<p>\u8981\u83b7\u5f97\u597d\u7684\u96c6\u6210\u6548\u679c\uff0c\u90a3\u4e48\u6211\u4eec\u9700\u8981\u4e2a\u4f53\u7684\u5b66\u4e60\u5668\u62e5\u6709\u4e00\u5b9a\u7684\u51c6\u786e\u6027\uff0c\u5e76\u4e14\u5404\u81ea\u4e4b\u524d\u8981\u5b58\u5728\u4e00\u5b9a\u7684\u5dee\u5f02<\/p>\n<ul>\n<li>\u51c6\u786e\u6027\uff1a\u5f53\u5b66\u4e60\u5668\u51c6\u786e\u6027\u8fc7\u4f4e\u65f6\uff0c\u51fa\u73b0\u9519\u8bef\u9884\u6d4b\u7684\u6982\u7387\u6bd4\u8f83\u5927\uff0c\u96c6\u6210\u540e\u56e0\u4e3a\u5c11\u6570\u670d\u4ece\u591a\u6570\uff0c\u8f83\u591a\u7684\u9519\u8bef\u4f1a\u8986\u76d6\u8f83\u5c11\u7684\u6b63\u786e\uff0c\u6700\u7ec8\u5bfc\u81f4\u6574\u4f53\u7684\u9519\u8bef\uff0c\u4e5f\u5c31\u662f\u8bf4\uff0c\u4f1a\u5bfc\u81f4\u6574\u4f53\u7684\u51c6\u786e\u6027\u964d\u4f4e<\/li>\n<li>\u5dee\u5f02\u6027\uff1a\u5f53\u5b66\u4e60\u5668\u4e4b\u95f4\u5dee\u5f02\u6027\u8fc7\u4f4e\u65f6\uff0c\u5b66\u4e60\u5668\u5bf9\u4e8e\u6570\u636e\u7684\u9884\u6d4b\u7ed3\u679c\u76f8\u4f3c\uff0c\u9519\u8bef\u7684\u4e00\u8d77\u51fa\u9519\uff0c\u6b63\u786e\u7684\u4e00\u8d77\u6b63\u786e\uff0c\u65e0\u6cd5\u8fbe\u5230\u7528\u6b63\u786e\u7ea0\u6b63\u9519\u8bef\u7684\u6548\u679c\uff0c\u4e5f\u5c31\u662f\u8bf4\uff0c\u4e0d\u4f1a\u5bfc\u81f4\u6574\u4f53\u7684\u7f3a\u52e4\u7387\u63d0\u9ad8<\/li>\n<\/ul>\n<h3>\uff08\u4e8c\uff09Bagging<\/h3>\n<h4>1.Bagging\u6846\u67b6<\/h4>\n<p>Bagging\u662f\u901a\u8fc7\u6709\u653e\u56de\u7684\u62bd\u6837\u4ea7\u751f\u4e0d\u540c\u7684\u8bad\u7ec3\u96c6\uff0c\u4ece\u800c\u5f97\u5230\u6709\u5dee\u5f02\u7684\u5b66\u4e60\u5668\uff0c\u7136\u540e\u901a\u8fc7\u5c11\u6570\u670d\u4ece\u591a\u6570\u7684\u65b9\u5f0f\u5f97\u5230\u9884\u6d4b\u7ed3\u679c<\/p>\n<h4>2.booststrap\u62bd\u6837<\/h4>\n<p>\u5728\u6837\u672c\u96c6\u4e2d\u6709\u653e\u56de\u5730\u8fdb\u884c\u62bd\u6837\uff0c\u6bcf\u6b21\u62bd\u53d6\u7684\u6982\u7387\u76f8\u7b49\uff0c\u62bd\u4e2d\u7684\u6570\u636e\u4f5c\u4e3a\u8bad\u7ec3\u96c6\uff0c\u672a\u62bd\u4e2d\u7684\u6570\u636e\u4f5c\u4e3a\u6d4b\u8bd5\u96c6\uff0c\u5176\u4e2d\u8bad\u7ec3\u96c6\u5927\u7ea6\u536063.2\n<h4>3.\u6027\u80fd<\/h4>\n<ul>\n<li>Bagging\u4e0d\u4ec5\u80fd\u9002\u7528\u4e8e\u4e8c\u5206\u7c7b\uff0c\u800c\u4e14\u80fd\u4e0d\u7ecf\u4fee\u6539\u5730\u7528\u4e8e\u591a\u5206\u7c7b\u3001\u56de\u5f52\u4efb\u52a1<\/li>\n<li>Bagging\u4e3b\u8981\u5173\u6ce8\u964d\u4f4e\u65b9\u5dee\uff0c\u56e0\u6b64\u5176\u5728\u4e0d\u526a\u679d\u51b3\u7b56\u6811\u3001\u795e\u7ecf\u7f51\u7edc\u7b49\u6613\u53d7\u6837\u672c\u6270\u52a8\u5f71\u54cd\u7684\u5b66\u4e60\u5668\u4e0a\u6548\u679c\u660e\u663e<\/li>\n<\/ul>\n<h4>4.\u7b97\u6cd5<\/h4>\n<ul>\n<li>\u4ece\u6837\u672c\u96c6\u8fdb\u884cbooststrap\u62bd\u6837\uff0c\u5f97\u5230\u82e5\u5e72\u8bad\u7ec3\u96c6\u548c\u6d4b\u8bd5\u96c6<\/li>\n<li>\u5bf9\u6bcf\u4e2a\u8bad\u7ec3\u96c6\u548c\u6d4b\u8bd5\u96c6\u5206\u522b\u8bad\u7ec3\u6a21\u578b<\/li>\n<li>\u6700\u7ec8\u6bcf\u4e2a\u6a21\u578b\u6309\u4e00\u5b9a\u65b9\u5f0f\u8fdb\u884c\u7ec4\u5408\u5f97\u5230\u6700\u7ec8\u7684\u9884\u6d4b\u7ed3\u679c\n<ul>\n<li>\u5206\u7c7b\u4efb\u52a1\uff1a\u7b80\u5355\u6295\u7968\uff0c\u6bcf\u4e2a\u5b66\u4e60\u5668\u4e00\u7968\uff0c\u5c11\u6570\u670d\u4ece\u591a\u6570<\/li>\n<li>\u56de\u5f52\u4efb\u52a1\uff1a\u7b80\u5355\u5e73\u5747\uff0c\u6bcf\u4e2a\u5b66\u4e60\u5668\u7684\u503c\u52a0\u548c\u6c42\u5e73\u5747<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n<h4>5.\u968f\u673a\u68ee\u6797<\/h4>\n<ul>\n<li>\u968f\u673a\u68ee\u6797\u662f\u57fa\u4e8eBagging\u6846\u67b6\u5b9e\u73b0\u7684\u4e00\u79cd\u96c6\u6210\u5b66\u4e60\u7b97\u6cd5\uff0c\u5b83\u91c7\u7528\u51b3\u7b56\u6811\u6a21\u578b\u4f5c\u4e3a\u57fa\u5b66\u4e60\u5668<\/li>\n<li>API\uff1a\n<ul>\n<li><code>sklearn.ensemble.RandomForestClassifier()<\/code><\/li>\n<li><code>n_estimators<\/code>\uff1a\u51b3\u7b56\u6811\u7684\u6570\u91cf<\/li>\n<li><code>Criterion<\/code>\uff1a\u51b3\u7b56\u6811\u5212\u5206\u4f9d\u636e\uff08entropy\u6216\u8005gini[\u9ed8\u8ba4]\uff09<\/li>\n<li><code>max_depth<\/code>\uff1a\u6811\u7684\u6700\u5927\u6df1\u5ea6<\/li>\n<li><code>max_features=&quot;auto&quot;<\/code>\uff1a\u51b3\u7b56\u6811\u6784\u5efa\u65f6\u4f7f\u7528\u7684\u6700\u5927\u7279\u5f81\u6570\u91cf\n<ul>\n<li><code>&quot;auto&quot;<\/code>\u6216<code>&quot;sqrt&quot;<\/code>\uff1aauto\u9ed8\u8ba4\u91c7\u7528\u7684\u5c31\u662fsqrt\uff0c\u7ed3\u679c\u4e3a<code>sqrt(n_features)<\/code><\/li>\n<li><code>&quot;log2&quot;<\/code>\uff1a\u7ed3\u679c\u4e3a<code>log2(n_features)<\/code><\/li>\n<li><code>&quot;None&quot;<\/code>\uff1a\u7ed3\u679c\u4e3a<code>n_features<\/code><\/li>\n<\/ul>\n<\/li>\n<li><code>booststrap<\/code>\uff1a\u662f\u5426\u91c7\u7528\u6709\u653e\u56de\u62bd\u6837\uff0c\u9ed8\u8ba4\u4e3aTrue<\/li>\n<li><code>min_samples_split<\/code>\uff1a\u7ed3\u70b9\u5206\u88c2\u6240\u9700\u7684\u6700\u5c0f\u6837\u672c\u6570\uff0c\u9ed8\u8ba4\u4e3a2<\/li>\n<li><code>min_samples_leaf<\/code>\uff1a\u53f6\u5b50\u8282\u70b9\u7684\u6700\u5c0f\u6837\u672c\u6570\uff0c\u9ed8\u8ba4\u4e3a1<\/li>\n<li><code>min_impurity_split<\/code>\uff1a\u8282\u70b9\u5212\u5206\u7684\u6700\u5c0f\u4e0d\u7eaf\u5ea6\uff08\u57fa\u5c3c\u7cfb\u6570\u3001\u5747\u65b9\u5dee\u7b49\uff09<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n<h3>\uff08\u4e09\uff09Boosting<\/h3>\n<h4>1.Boosting\u6846\u67b6<\/h4>\n<p>\u6bcf\u4e00\u4e2a\u5b66\u4e60\u5668\u91cd\u70b9\u5173\u6ce8\u524d\u4e00\u4e2a\u5b66\u4e60\u5668\u4e0d\u8db3\u7684\u5730\u65b9\u8fdb\u884c\u8bad\u7ec3\uff0c\u901a\u8fc7\u52a0\u6743\u6295\u7968\u5f97\u5230\u6700\u7ec8\u7ed3\u679c<\/p>\n<h4>2.\u7b97\u6cd5<\/h4>\n<ul>\n<li>\u5148\u4ece\u521d\u59cb\u8bad\u7ec3\u96c6\u8bad\u7ec3\u4e00\u4e2a\u57fa\u5b66\u4e60\u5668<\/li>\n<li>\u6839\u636e\u57fa\u5b66\u4e60\u5668\u7684\u8868\u73b0\u5bf9\u8bad\u7ec3\u6837\u672c\u5206\u5e03\u8fdb\u884c\u8c03\u6574\uff0c\u4f7f\u5f97\u5148\u524d\u57fa\u5b66\u4e60\u5668\u505a\u9519\u7684\u8bad\u7ec3\u6837\u672c\u5728\u540e\u7eed\u5f97\u5230\u6700\u5927\u7684\u5173\u6ce8<\/li>\n<li>\u57fa\u4e8e\u8c03\u6574\u8fc7\u540e\u7684\u6837\u672c\u5206\u5e03\u8bad\u7ec3\u4e0b\u4e00\u4e2a\u57fa\u5b66\u4e60\u5668<\/li>\n<li>\u5982\u6b64\u53cd\u590d\uff0c\u6700\u7ec8\u5c06\u6240\u6709\u57fa\u5b66\u4e60\u5668\u8fdb\u884c\u52a0\u6743\u7ec4\u5408<\/li>\n<\/ul>\n<h3>\uff08\u56db\uff09AdaBoost<\/h3>\n<h4>1.\u7b80\u4ecb<\/h4>\n<ul>\n<li>\n<p>AdaBoost\u662fBoosting\u7b97\u6cd5\u7684\u4e00\u79cd\u5b9e\u73b0\uff0c\u7528\u4e8e\u5206\u7c7b\u95ee\u9898<\/p>\n<\/li>\n<li>\n<p>AdaBoost\u901a\u8fc7\u5f31\u5206\u7c7b\u5668\u7684\u7ebf\u6027\u7ec4\u5408\u6784\u9020\u5f3a\u5206\u7c7b\u5668<\/p>\n<\/li>\n<li>\n<p>\u8bad\u7ec3\u65f6\uff0c\u8d4b\u4e88\u6837\u672c\u6743\u91cd\uff0c\u5e76\u5728\u8bad\u7ec3\u8fc7\u7a0b\u4e2d\u52a8\u6001\u8c03\u6574\uff0c\u5206\u9519\u7684\u6837\u672c\u4f1a\u589e\u52a0\u6743\u91cd<\/p>\n<\/li>\n<\/ul>\n<h4>2.\u6b65\u9aa4<\/h4>\n<ul>\n<li>\u5047\u8bbe\u5b58\u5728N\u4e2a\u6837\u672c\uff0c\u6bcf\u4e2a\u6837\u672c\u8d4b\u4e88\u6743\u91cd\u4e3a1\/N<\/li>\n<li>\u5728\u8bad\u7ec3\u4e2d\uff0c\u82e5\u6837\u672c\u5206\u7c7b\u9519\u8bef\uff0c\u5219\u63d0\u5347\u5176\u6743\u91cd\uff1b\u5206\u7c7b\u6b63\u786e\uff0c\u5219\u964d\u4f4e\u5176\u6743\u91cd<\/li>\n<li>\u6743\u91cd\u66f4\u65b0\u8fc7\u540e\u7684\u6837\u672c\u4f5c\u4e3a\u4e0b\u4e00\u4e2a\u8bad\u7ec3\u5668\u7684\u8bad\u7ec3\u96c6<\/li>\n<li>\u6700\u7ec8\u52a0\u6743\u7ec4\u5408\u5f31\u5206\u7c7b\u5668\uff0c\u5206\u7c7b\u8bef\u5dee\u5c0f\u7684\u6743\u91cd\u5927\uff0c\u5206\u7c7b\u8bef\u5dee\u5927\u7684\u6743\u91cd\u5c0f<\/li>\n<\/ul>\n<h4>3.\u516c\u5f0f<\/h4>\n<ul>\n<li>\n<p>\u6700\u7ec8\u9884\u6d4b\u516c\u5f0f\uff1a\n$$\nH(x)=sign(\\sum_{i=1}^m \\alpha_ih_i(x))\n$$<\/p>\n<ul>\n<li>$ \\alpha $\uff1a\u4e3a\u6a21\u578b\u7684\u6743\u91cd<\/li>\n<li>$ m $\uff1a\u4e3a\u5f31\u5b66\u4e60\u5668\u7684\u6570\u91cf<\/li>\n<li>$ h_i(x) $\uff1a\u4e3a\u5f31\u5b66\u4e60\u5668<\/li>\n<li>$ H(x) $\uff1a\u4e3a\u5206\u7c7b\u7ed3\u679c\uff0c\u5927\u4e8e0\u4e3a\u6b63\u7c7b\uff0c\u5c0f\u4e8e0\u4e3a\u8d1f\u7c7b<\/li>\n<\/ul>\n<\/li>\n<li>\n<p>\u6a21\u578b\u6743\u91cd\u66f4\u65b0\u516c\u5f0f\uff1a\n$$\n\\alpha_t=\\frac12ln(\\frac{1-\\epsilon_t}{\\epsilon_t})\n$$<\/p>\n<ul>\n<li>$ \\epsilon_t $\uff1a\u8868\u793a\u7b2ct\u4e2a\u5f31\u5b66\u4e60\u5668\u7684\u9519\u8bef\u7387<\/li>\n<\/ul>\n<\/li>\n<li>\n<p>\u6837\u672c\u6743\u91cd\u66f4\u65b0\u516c\u5f0f\uff1a\n$$\nD_{t+1}=\\frac{D_t(x)}{Z_t}*\\begin{cases}e^{-\\alpha_t},\u9884\u6d4b\u503c=\u771f\u5b9e\u503c\\e^\\alpha_t,\u9884\u6d4b\u503c\\ne\u771f\u5b9e\u503c\\end{cases}\n$$<\/p>\n<ul>\n<li>$ Z_t $\uff1a\u5f52\u4e00\u5316\u503c\uff0c\u5373\u6240\u6709\u6837\u672c\u6743\u91cd\u7684\u603b\u548c<\/li>\n<li>$ D_t(x) $\uff1a\u4e3a\u6837\u672c\u6743\u91cd<\/li>\n<li>$ \\alpha_t $\uff1a\u6a21\u578b\u6743\u91cd<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n<h3>\uff08\u4e94\uff09GBDT<\/h3>\n<h4>1.\u63d0\u5347\u6811<\/h4>\n<ul>\n<li>\u6b65\u9aa4\n<ul>\n<li>\u521d\u59cb\u5316$ f_0(x)=0 $<\/li>\n<li>\u5bf9\u4e8e$ m=1,2,\u2026,M $\n<ul>\n<li>\u8ba1\u7b97\u6b8b\u5dee\uff1a$ r_{mi}=y_i-f_{m-1}(x),i=1,2,\u2026,N $<\/li>\n<li>\u62df\u5408\u6b8b\u5dee$ r_{mi} $\u5f97\u5230\u56de\u5f52\u6811$ h_m(x) $<\/li>\n<li>\u66f4\u65b0$ f_m(x)=f_{m-1}(x)+h_m(x) $<\/li>\n<\/ul>\n<\/li>\n<li>\u5f97\u5230\u56de\u5f52\u95ee\u9898\u63d0\u5347\u6811\uff1a$ f_M(x)=\\sum_{m=1}^M h_m(x) $<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n<h4>2.\u68af\u5ea6\u63d0\u5347\u6811<\/h4>\n<ul>\n<li>\u68af\u5ea6\u63d0\u5347\u6811\u5229\u7528\u635f\u5931\u51fd\u6570\u7684\u8d1f\u68af\u5ea6\u4f5c\u4e3a\u63d0\u5347\u6811\u7b97\u6cd5\u4e2d\u6b8b\u5dee\u7684\u8fd1\u4f3c\u503c\uff0c\u5b9e\u9645\u4e0a\u4e00\u822c\u8d1f\u68af\u5ea6\u7b49\u4e8e\u6b8b\u5dee<\/li>\n<\/ul>\n<h4>3.GBDT\u7b97\u6cd5<\/h4>\n<ul>\n<li>\n<p>\u521d\u59cb\u5316\u5f31\u5b66\u4e60\u5668\uff1a\u5bf9\u635f\u5931\u51fd\u6570\u6c42\u6700\u5c0f\u503c\n$$\nf_0(x)=arg\\space min_c\\sum_{i=1}^N L(y_i,c)\n$$<\/p>\n<\/li>\n<li>\n<p>\u5bf9\u4e8e$ m=1,2,\u2026,M $\uff1a<\/p>\n<ul>\n<li>\n<p>\u5bf9\u6bcf\u4e00\u4e2a\u6837\u672ci,$ i=1,2,\u2026,N $\u6c42\u8d1f\u68af\u5ea6\uff1a\n$$\nr_{im}=-[\\frac{\\partial L(y_i,f(x_i))}{\\partial f(x_i)}]f(x)=f_{m-1}(x)\n$$<\/p>\n<\/li>\n<li>\n<p>\u6839\u636e\u5212\u5206\u70b9\uff0c\u5c06\u6837\u672c\u5206\u4e3a\u4e24\u90e8\u5206\uff0c\u6bcf\u4e00\u90e8\u5206\u4f7f\u7528\u5176\u8d1f\u68af\u5ea6\u7684\u5e73\u5747\u503c\u4f5c\u4e3a\u9884\u6d4b\u503c<\/p>\n<\/li>\n<li>\n<p>\u5c06\u6bcf\u4e00\u4e2a\u6837\u672c\u7684\u8d1f\u68af\u5ea6\u4f5c\u4e3a\u8be5\u6837\u672c\u7684\u76ee\u6807\u503c\uff0c\u8be5\u6837\u672c\u6240\u5c5e\u90e8\u5206\u7684\u8d1f\u68af\u5ea6\u5e73\u5747\u503c\u4f5c\u4e3a\u9884\u6d4b\u503c<\/p>\n<\/li>\n<li>\n<p>\u7ee7\u7eed\u8ba1\u7b97\u8fd9\u4e00\u8f6e\u6837\u672c\u7684\u8d1f\u68af\u5ea6\uff0c\u5982\u6b64\u5f80\u590d<\/p>\n<\/li>\n<\/ul>\n<\/li>\n<li>\n<p>\u6700\u7ec8\u4f7f\u7528\u5f3a\u5b66\u4e60\u5668\u7684\u9884\u6d4b\u7ed3\u679c\u4e0e\u6240\u6709\u5f31\u5b66\u4e60\u5668\u7684\u9884\u6d4b\u7ed3\u679c\u76f8\u52a0\u5f97\u5230\u6700\u7ec8\u7684\u9884\u6d4b\u7ed3\u679c<\/p>\n<\/li>\n<\/ul>\n<h3>\uff08\u516d\uff09\u96c6\u6210\u7b97\u6cd5\u591a\u6837\u6027<\/h3>\n<h4>1.\u6570\u636e\u6837\u672c\u6270\u52a8<\/h4>\n<p>\u5bf9\u4e8e\u7ed9\u5b9a\u7684\u6570\u636e\u96c6\u901a\u8fc7\u91c7\u6837\u6cd5\u4ece\u4e2d\u5f97\u5230\u5177\u6709\u5dee\u5f02\u7684\u6570\u636e\u5b50\u96c6<\/p>\n<h4>2.\u8f93\u5165\u5c5e\u6027\u6270\u52a8<\/h4>\n<p>\u8bad\u7ec3\u6837\u672c\u901a\u5e38\u7531\u4e00\u7ec4\u5c5e\u6027\u8fdb\u884c\u63cf\u8ff0\uff0c\u53ef\u4ee5\u57fa\u4e8e\u8fd9\u4e9b\u5c5e\u6027\u7684\u4e0d\u540c\u7ec4\u5408\u4ea7\u751f\u4e0d\u540c\u7684\u6570\u636e\u5b50\u96c6<\/p>\n<h4>3.\u7b97\u6cd5\u53c2\u6570\u6270\u52a8<\/h4>\n<p>\u901a\u8fc7\u968f\u673a\u8bbe\u7f6e\u4e0d\u540c\u7684\u53c2\u6570\uff0c\u5bf9\u6a21\u578b\u53c2\u6570\u52a0\u5165\u5c0f\u8303\u56f4\u7684\u968f\u673a\u6270\u52a8\uff0c\u4ece\u800c\u4ea7\u751f\u5dee\u522b\u8f83\u5927\u7684\u5b66\u4e60\u5668<\/p>\n<h3>\uff08\u4e03\uff09XGBoost<\/h3>\n<h4>1.XGBoost\u539f\u7406<\/h4>\n<p>XGBoost\u662f\u5bf9\u68af\u5ea6\u63d0\u5347\u7b97\u6cd5\u7684\u6539\u8fdb\uff1a<\/p>\n<ul>\n<li>\u6c42\u89e3\u635f\u5931\u51fd\u6570\u65f6\u4f7f\u7528\u6cf0\u52d2\u4e8c\u9636\u5c55\u5f00<\/li>\n<li>\u5728\u635f\u5931\u51fd\u6570\u4e2d\u52a0\u5165\u4e86\u6b63\u5219\u5316\u9879<\/li>\n<li>\u4ece\u635f\u5931\u51fd\u6570\u4e2d\u63a8\u5bfc\u51fa\u4e00\u4e2a\u6811\u8282\u70b9\u5206\u88c2\u6307\u6807<\/li>\n<\/ul>\n<p>\u5bf9\u4e8e\u6784\u5efa\u6700\u4f18\u6a21\u578b\uff0c\u4e00\u822c\u91c7\u7528\u7684\u65b9\u6cd5\u662f\u6700\u5c0f\u5316\u8bad\u7ec3\u6570\u636e\u7684\u635f\u5931\u51fd\u6570\uff1a<\/p>\n<ul>\n<li>\n<p>\u7ecf\u9a8c\u98ce\u9669\u6700\u5c0f\u5316\uff1a\n$$\nmin\\frac1N\\sum_{i=1}^NL(y_i,f(x_i))\n$$<\/p>\n<\/li>\n<li>\n<p>\u7ed3\u6784\u98ce\u9669\u6700\u5c0f\u5316\uff1a\n$$\nmin\\frac1N\\sum_{i=1}^NL(y_i,f(x_i))+\\Omega(f)\n$$<\/p>\n<\/li>\n<\/ul>\n<h4>2.XGBoost\u516c\u5f0f<\/h4>\n<p>$$\nobj(\\theta)=\\sum_i^n L(y_i,\\hat y_i)+\\sum_{k=1}^K\\Omega(f_k)\n$$<\/p>\n<ul>\n<li>\n<p>\u7b2c\u4e00\u9879\uff1a\u8868\u793a\u5f3a\u5b66\u4e60\u5668\u7684\u635f\u5931<\/p>\n<\/li>\n<li>\n<p>\u7b2c\u4e8c\u9879\uff1a\u8868\u793aK\u4e2a\u5f31\u5b66\u4e60\u5668\u7684\u590d\u6742\u5ea6<\/p>\n<\/li>\n<\/ul>\n<h4>3.\u5f31\u5b66\u4e60\u5668\u590d\u6742\u5ea6<\/h4>\n<p>$$\n\\Omega(f)=\\gamma T+\\frac12\\lambda||\\omega||^2\n$$<\/p>\n<ul>\n<li>\u7b2c\u4e00\u9879\uff1a$T$\u8868\u793a\u4e00\u68f5\u6811\u7684\u53f6\u5b50\u7ed3\u70b9\u7684\u6570\u91cf\uff0c$\\gamma$\u8868\u793a\u8be5\u9879\u7684\u8c03\u8282\u7cfb\u6570<\/li>\n<li>\u7b2c\u4e8c\u9879\uff1a$\\omega$\u8868\u793a\u53f6\u5b50\u7ed3\u70b9\u8f93\u51fa\u503c\u7ec4\u6210\u7684\u5411\u91cf\uff0c$\\lambda$\u8868\u793a\u8be5\u9879\u7684\u8c03\u8282\u7cfb\u6570<\/li>\n<\/ul>\n<h4>4.\u6253\u5206\u51fd\u6570<\/h4>\n<p>$$\nobj=-\\frac12\\sum_{i=1}^T(\\frac{G_i^2}{H_i+\\lambda})+\\gamma T\n$$<\/p>\n<ul>\n<li>\u4f7f\u7528\u65b9\u6cd5\uff1a\n<ul>\n<li>\u6211\u4eec\u5bf9\u4e8e\u4e00\u4e2a\u8282\u70b9\uff0c\u5206\u88c2\u524d\u6c42\u6253\u5206\u51fd\u6570\uff0c\u5206\u88c2\u540e\u6c42\u4e24\u4e2a\u8282\u70b9\u7684\u6253\u5206\u51fd\u6570\u4e4b\u548c<\/li>\n<li>\u8ba1\u7b97\u5206\u88c2\u524d\u51cf\u53bb\u5206\u88c2\u540e\u7684\u5206\u6570\n<ul>\n<li>\u82e5\u5927\u4e8e0\uff0c\u8bf4\u660e\u5206\u88c2\u540e\u6811\u7684\u7ed3\u6784\u635f\u5931\u66f4\u5c0f\uff0c\u8003\u8651\u5206\u88c2<\/li>\n<li>\u82e5\u5c0f\u4e8e0\uff0c\u8bf4\u660e\u5206\u88c2\u540e\u6811\u7684\u7ed3\u6784\u635f\u5931\u66f4\u5927\uff0c\u907f\u514d\u5206\u88c2<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n<h4>5.\u5e38\u7528API<\/h4>\n<ul>\n<li>\n<p>\u901a\u7528\u53c2\u6570<\/p>\n<ul>\n<li>\n<p><code>booster<\/code>\uff1a\u9ed8\u8ba4\u4f7f\u7528<code>gbtree<\/code><\/p>\n<ul>\n<li><code>gbtree<\/code>\uff1a\u4f7f\u7528\u6811\u6a21\u578b<\/li>\n<li><code>gblinear<\/code>\uff1a\u4f7f\u7528\u7ebf\u6027\u6a21\u578b<\/li>\n<\/ul>\n<\/li>\n<li>\n<p><code>silent<\/code>\uff1a\u9ed8\u8ba4\u4f7f\u7528<code>0<\/code><\/p>\n<ul>\n<li><code>0<\/code>\uff1a\u6253\u5370\u8fd0\u884c\u4fe1\u606f<\/li>\n<li><code>1<\/code>\uff1a\u4e0d\u6253\u5370\u8fd0\u884c\u4fe1\u606f<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n<\/li>\n<li>\n<p>TreeBooster<\/p>\n<ul>\n<li><code>eta<\/code>\uff1a\u5b66\u4e60\u7387\uff0c\u9ed8\u8ba4\u4e3a<code>0.3<\/code><\/li>\n<li><code>gamma<\/code>\uff1a\u8282\u70b9\u5206\u88c2\u6240\u9700\u7684\u6700\u5c0f\u635f\u5931\u51fd\u6570\u4e0b\u964d\u503c\uff0c\u9ed8\u8ba4\u4e3a<code>0<\/code><\/li>\n<li><code>max_depth<\/code>\uff1a\u6811\u7684\u6700\u5927\u6df1\u5ea6\uff0c\u9ed8\u8ba4\u4e3a<code>6<\/code><\/li>\n<li><code>min_child_weight<\/code>\uff1a\u9ed8\u8ba4\u4e3a<code>1<\/code>\n<ul>\n<li>\u5f53\u503c\u8fc7\u5927\u65f6\uff0c\u53ef\u4ee5\u907f\u514d\u6a21\u578b\u5b66\u4e60\u5230\u5c40\u90e8\u7684\u7279\u6b8a\u6837\u672c\uff0c\u4f46\u503c\u8fc7\u9ad8\u6613\u5bfc\u81f4\u6b20\u62df\u5408\uff0c<\/li>\n<\/ul>\n<\/li>\n<li><code>subsample<\/code>\uff1a\u63a7\u5236\u968f\u673a\u91c7\u6837\u7684\u6bd4\u4f8b\uff0c\u53d6\u503c\u8303\u56f4(0,1]\uff0c\u9ed8\u8ba4\u4e3a<code>1<\/code>\n<ul>\n<li>\u503c\u8fc7\u5927\uff1a\u6613\u5bfc\u81f4\u8fc7\u62df\u5408<\/li>\n<li>\u503c\u8fc7\u5c0f\uff1a\u6613\u5bfc\u81f4\u6b20\u62df\u5408<\/li>\n<li>\u5e38\u7528\u503c\uff1a0.5-1\uff0c0.5\u8868\u793a\u5e73\u5747\u91c7\u6837<\/li>\n<\/ul>\n<\/li>\n<li><code>colsample_bytree<\/code>\uff1a\u63a7\u5236\u6bcf\u68f5\u6811\u7279\u5f81\u91c7\u6837\u7684\u6bd4\u4f8b\uff0c\u9ed8\u8ba4\u4e3a<code>1<\/code>\uff0c\u5e38\u75280.5-1<\/li>\n<li><code>colsample_bylevel<\/code>\uff1a\u63a7\u5236\u6811\u6bcf\u4e00\u6b21\u5206\u88c2\u5bf9\u7279\u5f81\u91c7\u6837\u7684\u6bd4\u4f8b\uff0c\u9ed8\u8ba4\u4e3a<code>1<\/code><\/li>\n<li><code>alpha<\/code>\uff1a\u6743\u91cd\u7684L1\u6b63\u5219\u5316\u9879\uff0c\u9ed8\u8ba4\u4e3a<code>0<\/code><\/li>\n<li><code>scale_pos_weight<\/code>\uff1a\u9ed8\u8ba4\u4e3a<code>1<\/code>\n<ul>\n<li>\u5f53\u6837\u672c\u5341\u5206\u4e0d\u5e73\u8861\u65f6\uff0c\u53ef\u4ee5\u5c06\u8bcd\u6b64\u53c2\u6570\u8bbe\u5b9a\u4e3a\u4e00\u4e2a\u6b63\u503c\uff0c\u4f7f\u7b97\u6cd5\u66f4\u5feb\u6536\u655b<\/li>\n<li>\u901a\u5e38\u8bbe\u7f6e\u4e3a\u8d1f\u6837\u672c\u7684\u6570\u76ee\u4e0e\u6b63\u6837\u672c\u7684\u6570\u76ee\u7684\u6bd4\u503c<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n<\/li>\n<li>\n<p>LinearBooster<\/p>\n<ul>\n<li><code>lambda<\/code>\uff1a\u9ed8\u8ba4\u4e3a<code>0<\/code>\n<ul>\n<li>L2\u6b63\u5219\u5316\u60e9\u7f5a\u7cfb\u6570\uff0c\u589e\u52a0\u8be5\u503c\u4f1a\u5bfc\u81f4\u6a21\u578b\u66f4\u52a0\u4fdd\u5b88<\/li>\n<\/ul>\n<\/li>\n<li><code>alpha<\/code>\uff1a\u9ed8\u8ba4\u4e3a<code>0<\/code>\n<ul>\n<li>L1\u6b63\u5219\u5316\u60e9\u7f5a\u7cfb\u6570\uff0c\u589e\u52a0\u8be5\u503c\u4f1a\u5bfc\u81f4\u6a21\u578b\u66f4\u52a0\u4fdd\u5b88<\/li>\n<\/ul>\n<\/li>\n<li><code>lambda_bias<\/code>\uff1a\u9ed8\u8ba4\u4e3a<code>0<\/code>\n<ul>\n<li>\u504f\u7f6e\u4e0a\u7684L2\u6b63\u5219\u5316<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n<\/li>\n<li>\n<p>\u5b66\u4e60\u76ee\u6807\u53c2\u6570<\/p>\n<ul>\n<li><code>objective<\/code>\uff1a\u9ed8\u8ba4\u4e3a<code>reg:linear<\/code>\n<ul>\n<li><code>reg:linear<\/code>\uff1a\u7ebf\u6027\u56de\u5f52<\/li>\n<li><code>reg:logistic<\/code>\uff1a\u903b\u8f91\u56de\u5f52<\/li>\n<li><code>binary:logistic<\/code>\uff1a\u4e8c\u5206\u7c7b\u903b\u8f91\u56de\u5f52<\/li>\n<\/ul>\n<\/li>\n<li><code>eval_metric<\/code>\uff1a\u6307\u5b9a\u9a8c\u8bc1\u96c6\u7684\u6307\u6807\uff0c\u9ed8\u8ba4\u662f\u901a\u8fc7\u76ee\u6807\u51fd\u6570\u9009\u62e9\n<ul>\n<li><code>rmse<\/code>\uff1a\u5747\u65b9\u6839\u8bef\u5dee<\/li>\n<li><code>mae<\/code>\uff1a\u5e73\u5747\u7edd\u5bf9\u503c\u8bef\u5dee<\/li>\n<li><code>logloss<\/code>\uff1a\u8d1f\u5bf9\u6570\u4f3c\u7136\u51fd\u6570\u503c<\/li>\n<li><code>error<\/code>\uff1a\u9519\u8bef\u5206\u7c7b\u6570\u76ee\u4e0e\u5168\u90e8\u5206\u7c7b\u6570\u76ee\u6bd4\u503c<\/li>\n<li><code>auc<\/code>\uff1a\u66f2\u7ebf\u4e0b\u9762\u79ef<\/li>\n<\/ul>\n<\/li>\n<li><code>seed<\/code>\uff1a\u968f\u673a\u6570\u79cd\u5b50\uff0c\u9ed8\u8ba4\u4e3a<code>0<\/code><\/li>\n<\/ul>\n<\/li>\n<\/ul>\n<h2>\u9644\u5f55\uff1a<\/h2>\n<h4>1.\u4fdd\u5b58\u548c\u52a0\u8f7d\u6a21\u578b<\/h4>\n<pre><code class=\"language-python\">#\u4fdd\u5b58\njoblib.dump(\u6a21\u578b,\u4fdd\u5b58\u7684\u6587\u4ef6\u540d)\n\n#\u52a0\u8f7d\njoblib.load(\u6587\u4ef6\u540d)\n<\/code><\/pre>\n<h4>2.\u5b98\u65b9API\u53c2\u8003\u624b\u518c<\/h4>\n<p>\u5b98\u65b9API\u53c2\u8003\u624b\u518c\uff1ahttps:\/\/scikit-learn.org\/stable\/api\/index.html<\/p>\n<\/div>\n","protected":false},"excerpt":{"rendered":"","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"_jetpack_memberships_contains_paid_content":false,"footnotes":""},"categories":[2],"tags":[],"class_list":["post-77","post","type-post","status-publish","format-standard","hentry","category-2"],"jetpack_featured_media_url":"","jetpack_sharing_enabled":true,"_links":{"self":[{"href":"http:\/\/www.zxlearning.space\/index.php?rest_route=\/wp\/v2\/posts\/77","targetHints":{"allow":["GET"]}}],"collection":[{"href":"http:\/\/www.zxlearning.space\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"http:\/\/www.zxlearning.space\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"http:\/\/www.zxlearning.space\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"http:\/\/www.zxlearning.space\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=77"}],"version-history":[{"count":4,"href":"http:\/\/www.zxlearning.space\/index.php?rest_route=\/wp\/v2\/posts\/77\/revisions"}],"predecessor-version":[{"id":81,"href":"http:\/\/www.zxlearning.space\/index.php?rest_route=\/wp\/v2\/posts\/77\/revisions\/81"}],"wp:attachment":[{"href":"http:\/\/www.zxlearning.space\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=77"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"http:\/\/www.zxlearning.space\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=77"},{"taxonomy":"post_tag","embeddable":true,"href":"http:\/\/www.zxlearning.space\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=77"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}