{"id":25784,"date":"2023-10-03T17:30:00","date_gmt":"2023-10-03T09:30:00","guid":{"rendered":"https:\/\/tmrmds.co\/?p=25784"},"modified":"2023-10-03T17:08:13","modified_gmt":"2023-10-03T09:08:13","slug":"%e5%a6%82%e4%bd%95%e7%94%a8%e8%b3%87%e6%96%99%e7%a7%91%e5%ad%b8%e6%96%b9%e6%b3%95%e6%89%be%e5%87%ba%e9%ab%98cp%e5%80%bc%e7%94%a2%e5%93%81-%ef%bc%88%e9%99%84python%e7%a8%8b%e5%bc%8f%e7%a2%bc%ef%bc%89-4","status":"publish","type":"post","link":"https:\/\/tmrmds.co\/article-mds-operation\/25784\/","title":{"rendered":"\u5982\u4f55\u7528\u8cc7\u6599\u79d1\u5b78\u65b9\u6cd5\u627e\u51fa\u9ad8CP\u503c\u7522\u54c1 \uff08\u9644Python\u7a0b\u5f0f\u78bc\uff09"},"content":{"rendered":"\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t

\u5982\u4f55\u4f7f\u7528\u6a5f\u5668\u5b78\u7fd2\u63d0\u9ad8\u623f\u4ef2\u696d\u6f5b\u5728\u6210\u4ea4\u7387\uff1f
\u9032\u968e\u8cc7\u6599\u8655\u7406\u9762\u8207\u57fa\u790e\u5efa\u6a21(\u9644Python\u7a0b\u5f0f\u78bc)<\/h1>\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t

\u5148\u56de\u9867\u4e0a\u4e00\u7bc7\u6587\u7ae0\u5427\uff5e
\u5982\u4f55\u4f7f\u7528\u6a5f\u5668\u5b78\u7fd2\u63d0\u9ad8\u623f\u4ef2\u696d\u6f5b\u5728\u6210\u4ea4\u7387\uff1f\u8cc7\u6599\u8655\u7406\u9762\u57fa\u672c\u5fc3\u6cd5(\u9644Python\u7a0b\u5f0f\u78bc)<\/a><\/p><\/blockquote><\/div>

\u91dd\u5c0d\u524d\u4e00\u7bc7\u7684\u8cc7\u6599\u524d\u8655\u7406\uff0c\u6211\u5011\u5171\u53ef\u4ee5\u5f97\u523011\u500b\u7279\u5fb5\u8b8a\u6578(bathroom~price)\u548c1\u500b\u76ee\u6a19\u8b8a\u6578(interest_level)\uff0c\u5982\u57161<\/strong>\u6240\u793a\u3002<\/p>

\"\"
\u57161. \u7cfb\u52171_\u8cc7\u6599\u524d\u8655\u7406\u7d50\u679c<\/figcaption><\/figure>
<\/figcaption><\/figure>

\u9032\u968e\u8cc7\u6599\u8655\u7406<\/h3>

\u63a5\u4e0b\u4f86\uff0c\u6211\u5011\u5c07\u904b\u7528\u9032\u4e00\u6b65\u7684\u8cc7\u6599\u8655\u7406\u6280\u5de7\u4e26\u5efa\u69cb\u4e0d\u540c\u7684\u7684\u6a5f\u5668\u5b78\u7fd2\u6a21\u578b\uff0c\u6d41\u7a0b\u5982\u57162<\/strong>\u6240\u793a\u3002<\/p>

\"\"
\u57162. \u6a21\u578b\u5efa\u7acb\u6d41\u7a0b<\/figcaption><\/figure><\/figure>

\u5176\u4e2d\uff0c\u6211\u5011\u628a\u8cc7\u6599\u524d\u8655\u7406\u5206\u70ba<\/em>\u7279\u5fb5\u8655\u7406<\/em><\/strong>\u3001<\/em>\u8cc7\u6599\u5207\u5206<\/em><\/strong>\u3001<\/em>\u6a19\u6e96\u5316<\/em><\/strong>\u4e09\u6b65\u9a5f\uff0c\u540c\u6642\u4e5f\u6703\u628a\u64cd\u4f5c\u65b9\u6cd5\u53ca\u6210\u679c\u5448\u73fe\u65bc\u4e0b\u5217\u5167\u5bb9\u3002<\/em><\/p><\/blockquote>

\u5728\u9032\u5165\u5230\u4e0b\u4e00\u968e\u6bb5\u524d\uff0c\u9019\u908a\u9644\u4e0a\u4e86python\u7a0b\u5f0f\u78bc\u4f9b\u5927\u5bb6\u53c3\u8003\uff0c\u9023\u7d50\u5982\u4e0b\u3002
\u7a0b\u5f0f\u78bc\u4f86\u6e90:\u81fa\u7063\u884c\u92b7\u7814\u7a76Github<\/strong><\/a><\/p>

\u7279\u5fb5\u8655\u7406<\/h3>

\u4e3b\u8981\u91dd\u5c0d\u300cfeatures\u300d\u6b04\u4f4d\u88e1\u7684\u6240\u6709\u8cc7\u6599\uff0c\u9019\u908a\u53ef\u4ee5\u60f3\u50cf\u70bahashtag\u7684\u6982\u5ff5\uff0c\u5167\u5bb9\u591a\u70ba\u63cf\u8ff0\u623f\u5c4b\u7279\u5fb5\u7684\u5b57\u8a5e\uff0c\u9019\u7a2e\u5f62\u5f0f\u53ef\u4ee5\u8b93\u6d88\u8cbb\u8005\u5feb\u901f\u4e86\u89e3\u5230\u623f\u6771\u5c0d\u8a72\u623f\u5c4b\u6240\u6b32\u5448\u73fe\u7684\u91cd\u9ede\uff0c\u4f46\u6b64\u7a2e\u8cc7\u6599\u901a\u5e38\u6703\u7522\u751f\u5169\u500b\u5c0f\u554f\u984c<\/p>

1. \u6bcf\u7b46\u8cc7\u6599\u88e1\u53ef\u80fd\u6703\u5177\u6709\u4e0d\u540c\u7684hashtag
2. \u4e0d\u540c\u7b46\u8cc7\u6599\u4e2d\u53ef\u80fd\u6703\u542b\u6709\u76f8\u540c\u7684hashtag<\/p>

\u56e0\u6b64\uff0c\u70ba\u4e86\u6311\u9078\u51fa\u76f8\u5c0d\u91cd\u8981\u7684\u503c\uff0c\u6211\u5011\u5206\u5225\u8a08\u7b97\u6bcf\u500b\u5b57\u8a5e\u51fa\u73fe\u7684\u983b\u7387\uff0c\u4e26\u628a\u201d\u983b\u7387\u5927\u65bc100\u7684\u5b57\u8a5e\u201d\u7576\u4f5c\u7be9\u9078\u689d\u4ef6\u7522\u51fa\u6578\u7d44\u5b57\u8a5e\uff0c\u6700\u5f8c\u5c07\u7b26\u5408\u7be9\u9078\u689d\u4ef6\u7684\u5b57\u8a5e\u8f49\u63db\u70ba\u4e8c\u5143\u8b8a\u6578(dummy variable)\u7684\u5f62\u5f0f(\u517169\u500b)\u3002\u8209\u4f8b\u4f86\u8aaa\uff0c\u82e5\u6709\u4e00\u6b04\u4f4d\u70baElevator\uff0c\u5247\u5728\u300cfeatures\u300d\u7684\u6bcf\u7b46\u8cc7\u6599\u4e2d\u53ea\u8981\u51fa\u73feElevator\uff0c\u90a3\u9ebc\u5728Elevator\u7684\u6b04\u4f4d\u8655\u5c31\u586b\u4e0a1\uff0c\u53cd\u4e4b\u586b\u4e0a0\uff0c\u524d\u5f8c\u6bd4\u8f03\u5982\u57163<\/strong>\u6240\u793a\u3002<\/strong><\/p>

\"\"
\u57163. \u8f49\u63db\u4e8c\u5143\u8b8a\u6578\u6b04\u4f4d<\/figcaption><\/figure><\/figure>

\u64cd\u4f5c\u81f3\u6b64\uff0c\u4fbf\u53c8\u7522\u751f\u4e86\u65b0\u7684\u554f\u984c\uff0c\u7531\u65bc\u64c1\u6709\u904e\u591a\u7684\u6b04\u4f4d\uff0c\u5bb9\u6613\u8b93\u6574\u9ad4\u8b8a\u7570\u904e\u5927\u5c0e\u81f4\u6a21\u578b\u904e\u5ea6\u64ec\u5408(overfitting)\uff0c\u6545\u501f\u52a9\u4e86\u7279\u5fb5\u7be9\u9078\u7684\u5957\u4ef6(SelectFromModel)\uff0c\u4f86\u5e6b\u52a9\u6211\u5011\u89e3\u6c7a\u9019\u500b\u554f\u984c\uff0c\u4ecb\u7d39\u5982\u57164<\/strong>\u6240\u793a(\u5176\u4e2destimator\u9700\u586b\u5165\u7684\u5167\u5bb9\uff0c\u5c07\u7559\u81f3\u5e95\u4e0b\u57fa\u790e\u5efa\u6a21\u90e8\u5206\u518d\u9032\u884c\u4ecb\u7d39)\u3002<\/p>

\"\"
\u57164. SelectFromModel\u7c21\u4ecb<\/figcaption><\/figure><\/figure>

\u6709\u8208\u8da3\u7684\u8b80\u8005\u53ef\u4ee5\u53c3\u8003<\/em>\u6b64\u9023\u7d50<\/em><\/a>\uff0c\u5c07\u6709\u66f4\u8a73\u7d30\u7684\u539f\u6587\u4ecb\u7d39\u3002<\/em><\/p><\/blockquote>

\u8cc7\u6599\u5207\u5206<\/h3>

\u70ba\u5f8c\u7e8c\u5efa\u7acb\u6a5f\u5668\u5b78\u7fd2\u6a21\u578b\u7684\u4ea4\u53c9\u9a57\u8b49(Cross Validation)\uff0c\u4ee5\u4fbf\u5224\u5225\u8a72\u6a21\u578b\u6709\u7121overfitting\u548c\u6bd4\u8f03\u4e0d\u540c\u6a21\u578b\u9593\u4e86\u89e3\u4f55\u7a2e\u6a21\u578bunderfitting\uff0c\u5728\u6b64\u5c07\u6574\u500b\u8cc7\u6599\u5207\u5206\u62108:2\u7684\u5f62\u5f0f\u3002<\/p>

\u6a19\u6e96\u5316<\/h3>

\u78ba\u4fdd\u5efa\u7acb\u6a21\u578b\u6642\u8f03\u4e0d\u53d7\u539f\u59cb\u8cc7\u6599\u55ae\u4f4d\u4e0a\u7684\u5f71\u97ff\uff0c\u5c07\u8cc7\u6599\u4e2d\u7684\u6578\u503c\u6b04\u4f4d\u9032\u884c\u6a19\u6e96\u5316(\u6bcf\u7b46\u8cc7\u6599\u5148\u6263\u9664\u8a72\u6b04\u4f4d\u5e73\u5747\u518d\u9664\u4ee5\u8a72\u6b04\u4f4d\u4e4b\u6a19\u6e96\u5dee)\uff0c\u4f7f\u6700\u5f8c\u5404\u6b04\u4f4d\u7d0499%\u7684\u8cc7\u6599\u5747\u843d\u5728-3\u52303\u4e4b\u9593\uff0c\u9032\u884c\u5230\u6b64\u6b65\u9a5f\u6642\u9700\u975e\u5e38\u5c0f\u5fc3\uff0c\u4e0d\u53ef\u5c07\u524d\u9762\u7279\u5fb5\u8655\u7406\u4e4b\u4e8c\u5143\u8b8a\u6578\u6b04\u4f4d\u9032\u884c\u6a19\u6e96\u5316\uff0c\u5426\u5247\uff0c\u5c07\u6703\u5c0e\u81f4\u6a21\u578b\u7522\u751f\u504f\u8aa4\u3002<\/p>

\u6b64\u5916\uff0c\u6a19\u6e96\u5316\u6a21\u578b\u4e2d\u53ea\u6709 fit \u8a13\u7df4\u8cc7\u6599\u96c6(training set)\u800c\u975e\u9023\u540c\u6e2c\u8a66\u8cc7\u6599\u96c6(testing set)\u4e00\u540c\u9032\u884c\u7684\u539f\u56e0\u70ba\u78ba\u4fdd\u5169\u8005\u5e73\u5747\u6578\u548c\u6a19\u6e96\u5dee\u76f8\u540c\uff0c\u82e5 fit \u6e2c\u8a66\u8cc7\u6599\u96c6\u5c07\u6703\u7522\u751f\u4e0d\u540c\u7684\u5e73\u5747\u6578\u4ee5\u53ca\u8f03\u5927\u7684\u6a19\u6e96\u5dee\u5c0e\u81f4\u6a21\u578b\u7d50\u679c\u504f\u8aa4\u3002<\/em><\/p><\/blockquote>

\u9032\u968e\u8cc7\u6599\u8655\u7406\u5f8c\u7684\u6210\u679c\u6aa2\u8996<\/h3>

\u4ee5\u8a13\u7df4\u8cc7\u6599\u96c6\u70ba\u4f8b\uff0c\u5982\u57165<\/strong>\u6240\u793a\u3002(\u5e95\u4e0b\u70ba\u5f8c\u7e8c\u5efa\u7acbRandom Forest\u7684\u8cc7\u6599\u96c6\uff0c\u5f9e\u539f\u672c75\u500b\u8b8a\u6578\u4e2d\u7be9\u9078\u51fa6\u500b\u9023\u7e8c\u8b8a\u6578\u548c12\u500b\u985e\u5225\u8b8a\u6578)<\/p>

\"\"
\u57165. \u9032\u968e\u8cc7\u6599\u8655\u7406\u6210\u679c(\u4ee5RF\u70ba\u4f8b)<\/figcaption><\/figure><\/figure>

\u57fa\u790e\u5efa\u6a21<\/h3>

\u5b8c\u6210\u9032\u968e\u8cc7\u6599\u8655\u7406\u5f8c\uff0c\u4fbf\u53ef\u4ee5\u958b\u59cb\u9032\u884c\u5efa\u69cb\u5404\u500b\u6a5f\u5668\u5b78\u7fd2\u6a21\u578b\u4f86\u505a\u9810\u6e2c\uff0c\u800c\u6b64\u7bc7\u5373\u5c07\u64cd\u4f5c\u4e26\u4ecb\u7d39Random Forest\u3001XGBoost\u3001LightGBM \u4e09\u500b\u6a5f\u5668\u5b78\u7fd2\u6a21\u578b\u3002<\/p>

\u9996\u5148\uff0c\u9019\u908a\u5148\u91dd\u5c0d\u4e0a\u8ff0\u4e09\u500b\u6a5f\u5668\u5b78\u7fd2\u6a21\u578b\uff0c\u505a\u4e00\u500b\u7c21\u55ae\u7684\u6a21\u578b\u9078\u7528\u6bd4\u8f03\u4f9b\u5927\u5bb6\u53c3\u8003\u3002<\/p>

1. Random\u00a0Forest<\/h3>

\u76ee\u524d\u5341\u5206\u5ee3\u70ba\u4eba\u77e5\u4e14\u5165\u9580\u5bb9\u6613\u7684\u6a21\u578b\u4e4b\u4e00\uff0c\u900f\u904e\u96a8\u6a5f\u91cd\u8907\u53d6\u6a23\u7684\u65b9\u5f0f\u7522\u751f\u6a23\u672c\u96c6(bootstrap)\uff0c\u540c\u6642\u96a8\u6a5f\u5f9e\u7e3d\u7279\u5fb5\u6578\u4e2d\u9078\u64c7\u4e00\u5b9a\u6578\u91cf\u7684\u7279\u5fb5\u9032\u884c\u5206\u88c2\uff0c\u518d\u91cd\u8907\u591a\u6b21\u7522\u751f\u8a31\u591a\u9846\u6c7a\u7b56\u6a39\u7d44\u5408\u5728\u4e00\u8d77\uff0c\u7531\u4ed6\u5011\u9019\u4e9b\u591a\u6578\u7368\u7acb\u6295\u7968\u6c7a\u5b9a\u9810\u6e2c\u65b9\u5411\uff0c\u5f97\u51fa\u4e00\u500b\u5e73\u5747\u7684\u7d50\u679c\uff0c<\/strong>\u82e5\u60f3\u66f4\u6df1\u5165\u4e86\u89e3\uff0c\u53ef\u53c3\u8003\u6b64\u9023\u7d50<\/a>\u3002<\/p>

\u512a\u9ede:
(1)\u80fd\u8655\u7406\u591a\u7279\u5fb5\u7684\u8cc7\u6599
(2)\u7531\u65bc\u9032\u884c\u6c7a\u7b56\u7684\u65b9\u6cd5\u662f\u5f7c\u6b64\u7368\u7acb\u6295\u7968\uff0c\u8f03\u597d\u505a\u6210\u4e26\u884c\u65b9\u6cd5(\u64cd\u4f5c\u591a\u500b\u8655\u7406\u5668\u53bb\u89e3\u6c7a\u540c\u4e00\u500b\u554f\u984c)
(3)\u53ef\u4ee5\u900f\u904e\u5efa\u7acb\u591a\u9846\u6a39\u904b\u7528\u5927\u6578\u6cd5\u5247\u7684\u65b9\u5f0f\u907f\u514doverfitting
(4)\u5167\u5efa\u529f\u80fd\u53ef\u4ee5\u7be9\u9078\u51fa\u7279\u5fb5\u7684\u91cd\u8981\u7a0b\u5ea6<\/strong><\/p>

\u7f3a\u9ede:
(1)\u8f03\u4e0d\u63a8\u85a6\u4f7f\u7528\u65bc\u8ff4\u6b78\u554f\u984c\uff0c\u672c\u8eab\u6f14\u7b97\u6cd5\u5c31\u4e0d\u662f\u7d66\u51fa\u4e00\u500b\u9023\u7e8c\u7684\u8f38\u51fa\uff0c\u8f03\u5bb9\u6613overfitting
(2)\u8655\u7406\u566a\u97f3\u8f03\u5927\u7684\u554f\u984c\u6642\u5bb9\u6613overfitting<\/strong><\/p>

2. XGBoost(Extreme Gradient Boosting\u00a0)<\/h3>

\u76f8\u4fe1\u5927\u5bb6\u5c0d\u65bc\u9019\u500b\u90fd\u4e0d\u964c\u751f\uff0c\u6d3b\u7528\u65bc\u5404\u5927\u7af6\u8cfd\u4e2d\uff0c\u5728\u7f3a\u5931\u503c\u5f97\u8655\u7406\u53ca\u6574\u9ad4\u6e96\u78ba\u5ea6\u4e0a\u90fd\u975e\u5e38\u512a\u7570\uff0c\u9019\u908a\u5c31\u76f4\u63a5\u4ecb\u7d39\u4f7f\u7528\u7684\u512a\u7f3a\u9ede\uff0c\u7c21\u55ae\u4f86\u8aaa\uff0c\u6bcf\u6b21\u6dfb\u52a0\u4e00\u68f5\u6a39\u5c31\u662f\u589e\u52a0\u5b78\u7fd2\u4e00\u500b\u65b0\u51fd\u6578\uff0c\u900f\u904e\u64ec\u5408\u4e0a\u6b21\u7684\u932f\u8aa4\u7d93\u9a57\uff0c\u53bb\u63d0\u9ad8\u6574\u9ad4\u7684\u6e96\u78ba\u5ea6<\/strong>\uff0c\u82e5\u60f3\u9032\u4e00\u6b65\u4e86\u89e3\uff0c\u53ef\u53c3\u8003\u6b64\u9023\u7d50<\/a>\u3002<\/p>

\u512a\u9ede:
(1)\u52a0\u5165regularization\u907f\u514doverfitting
(2)\u900f\u904e\u7a00\u758f\u611f\u77e5\u6f14\u7b97\u6cd5\u8655\u7406\u7f3a\u5931\u503c\uff0c\u9047\u5230\u7f3a\u5931\u503c\u6642\u81ea\u52d5\u6b78\u985e\u5230\u9810\u8a2d\u65b9\u5411
(3)\u6548\u7387\u8f03\u4f73<\/strong><\/p>

\u7f3a\u9ede:
(1)\u9810\u6392\u5e8f\u5bb9\u6613\u82b1\u8cbb\u904e\u591a\u6642\u9593:<\/strong>\u96d6\u7136\u53ef\u4ee5\u900f\u904e\u524d\u8ff0\u964d\u4f4e\u5c0b\u627e\u6700\u4f73\u5206\u88c2\u9ede\u7684\u8a08\u7b97\u91cf\uff0c\u4f46\u904e\u7a0b\u4e2d\u4ecd\u9700\u8981\u904d\u6b77\u6240\u6709\u8cc7\u6599\u96c6<\/p>

3. LightGBM(Light Gradient Boosting\u00a0Machine)<\/h3>

\u6700\u5ee3\u70ba\u4eba\u77e5\u7684\u5730\u65b9\u5c31\u662f\u4ed6\u7684\u901f\u5ea6\u53ca\u4e0d\u5360\u7528\u904e\u591a\u8a18\u61b6\u9ad4\u7b49\u512a\u52e2\uff0c\u5728\u517c\u5177XGBoost\u7684\u6e96\u78ba\u5ea6\u4e0b\u52a0\u5feb\u6574\u9ad4model\u7684\u8a13\u7df4\u901f\u5ea6\u3002\u8207\u5176\u5728\u65b0\u589e\u5176\u4ed6\u6a39\u8449\u6642\u78ba\u5b9a\u6240\u6709\u5206\u652f\uff0c\u53ea\u78ba\u5b9a\u5176\u4e2d\u67d0\u90e8\u5206\u5728\u6642\u9593\u4e0a\u8f03\u70ba\u7bc0\u7701\uff0c\u6bcf\u6b21\u5206\u88c2\u6642\u5747\u6703\u4ee5\u80fd\u5920\u63d0\u4f9b\u6700\u5927\u5dee\u7570(\u5206\u88c2\u6548\u76ca\u6700\u5927)\u7684\u9ede\u51fa\u767c<\/strong>\uff0c\u82e5\u5927\u5bb6\u60f3\u7e7c\u7e8c\u5f80\u4e0b\u6df1\u5165\uff0c\u53ef\u53c3\u8003\u6b64\u9023\u7d50<\/a>\u3002<\/p>

\u512a\u9ede:
(1)\u901f\u5ea6\u66f4\u5feb
(2)\u5167\u5b58\u66f4\u5c0f<\/strong><\/p>

\u7f3a\u9ede:<\/strong>
(1)\u6709\u53ef\u80fd\u6703\u5c0e\u81f4overfitting:<\/strong>\u5206\u88c2\u65b9\u5f0f\u8207\u524d\u8ff0\u7684XGBoost\u8ddfRandom Forest\u4e0d\u540c(level-wise)\uff0c\u63a1\u7528leaf-wise\u7684\u65b9\u5f0f\u4f46\u6709\u53ef\u80fd\u5c0e\u81f4\u6574\u500b\u6a39\u9577\u5f97\u904e\u6df1\uff0c\u6545\u9700\u8981\u8a2d\u5b9amax_depth\u53bb\u63a7\u5236\u3002<\/p>

\u4ee5\u4e0a\u4e0d\u540c\u6a21\u578b\u7684\u512a\u7f3a\u9ede\u6574\u7406\u6bd4\u8f03\u5982\u4e0b\uff0c\u5982\u57166<\/strong>\u6240\u793a\u3002<\/p>

\"\"
\u57166. \u6a21\u578b\u9078\u7528\u6bd4\u8f03<\/figcaption><\/figure><\/figure>

\u4ecb\u7d39\u5b8c\u6a21\u578b\u9078\u7528\u5f8c\uff0c\u4fbf\u53ef\u5206\u5225\u5c07\u5404\u6a21\u578b\u4ee3\u5165\u7279\u5fb5\u7be9\u9078\u5957\u4ef6(SelectFromModel)\u4e2d\u7684estimator\u9032\u884c\u7be9\u9078\uff0c\u5982\u57167\uff0c<\/strong>\u4e26\u5b8c\u6210\u57fa\u790e\u5efa\u6a21\uff0c\u8a13\u7df4\u6210\u679c\u5982\u57168<\/strong>\u6240\u793a\u3002<\/p>

\"\"
\u57167. \u7d93\u904e\u7279\u5fb5\u7be9\u9078\u5957\u4ef6\u7be9\u9078\u5f8c\u5404\u6a21\u578b\u7522\u51fa\u4e4b\u7d50\u679c<\/figcaption><\/figure>
<\/figcaption><\/figure>
\"\"
\u57168. \u8a13\u7df4\u6210\u679c<\/figcaption><\/figure><\/figure>

\u4ee5\u57168<\/strong>\u8a13\u7df4\u6210\u679c\u800c\u8a00\uff0c\u6211\u5011\u4f7f\u7528\u5206\u985e\u5e38\u7528\u7684Log Loss\u4f86\u8a55\u5224\u6210\u6548\uff0cLog Loss\u7684\u8a08\u7b97\u516c\u5f0f\u70ba\u7269\u7406\u5b78\u5e38\u63d0\u53ca\u7684\u71b5\uff08Entropy\uff09\u6216\u7a31\u300c\u4e82\u5ea6\u300d\u3002\u82e5\u4e82\u5ea6\u8d8a\u4f4e\uff0c\u5247\u8868\u793a\u5206\u985e\u8d8a\u8da8\u65bc\u7cbe\u6e96\uff0c\u6240\u4ee5\u6211\u5011\u53ef\u4ee5\u5f97\u77e5\u82e5\u57168<\/strong>\u7684\u6578\u503c\u8d8a\u4f4e\uff0c\u5247\u4ee3\u8868\u5206\u985e\u8d8a\u7cbe\u6e96\u3002\u81f3\u65bc\u70ba\u4f55\u4f7f\u7528Log Loss\uff0c\u5c07\u6703\u4fdd\u7559\u81f3\u4e0b\u4e00\u7bc7\u8a73\u7d30\u4ecb\u7d39\u3002<\/p>

\u5f9e\u7d50\u679c\u767c\u73fe\uff0cRandom Forest\u6a21\u578b\u5728\u300c\u7d93\u904e\u7279\u5fb5\u7be9\u9078\u6e2c\u8a66\u8cc7\u6599\u96c6\uff08\u50c5\u8abf\u6574\u91cd\u8981\u7279\u5fb5\u503c\uff09\u300d\u7684Log Loss\u8f03\u300c\u672a\u7d93\u904e\u7279\u5fb5\u7be9\u9078\u6e2c\u8a66\u8cc7\u6599\u96c6\uff08\u539f\u8cc7\u6599\uff09\u300d\u9084\u8981\u5c0f\uff0c\u610f\u5473\u4e82\u5ea6\u8f03\u5c0f\uff0c\u4ee3\u8868\u300c\u7279\u5fb5\u7be9\u9078\u300d\u5728\u672c\u6a21\u578b\u662f\u6709\u4e00\u5b9a\u7684\u6548\u679c\u3002<\/p>

\u4e0d\u904e\u5728XGBoost\u53caLightGBM\u6a21\u578b\u5247\u662f\u4f7f\u7528\u300c\u672a\u7d93\u904e\u7279\u5fb5\u7be9\u9078\u6e2c\u8a66\u8cc7\u6599\u96c6\uff08\u539f\u8cc7\u6599\uff09\u300d\u7684\u4e82\u5ea6\u4f86\u5f97\u66f4\u5c0f\uff0c\u800c\u4e14\u6548\u679c\u6bd4Random Forest\u6a21\u578b\u9084\u8981\u4f86\u5f97\u66f4\u4f73\uff0c\u6240\u4ee5\u4f9d\u8cc7\u6599\u79d1\u5b78\u6a21\u578b\u8a55\u4f30\u6307\u6a19\u4f86\u8aaa\uff0c\u5efa\u8b70\u6211\u5011\u53ef\u4ee5\u9078\u64c7LightGBM\u6a21\u578b\u4f86\u9032\u884c\u5f8c\u7e8c\u76ee\u6a19\u8b8a\u6578(\u8a62\u554f\u5ea6)\u7684\u9810\u6e2c\u3002<\/p>

\u63a5\u4e0b\u4f86\u4fbf\u53ef\u4ee5\u67e5\u770b\u76ee\u6a19\u8b8a\u6578(\u8a62\u554f\u5ea6)\u7684\u5206\u985e\u60c5\u5f62\uff0c\u9019\u908a\u5c07\u6703\u4ee5LightGBM\u70ba\u4f8b\uff0c\u9810\u6e2c\u6210\u679c\u5982\u57169<\/strong>\u6240\u793a\u3002<\/p>

\"\"
\u57169. LGBM\u9810\u6e2c\u6210\u679c<\/figcaption><\/figure>
<\/figcaption><\/figure>

\u5f9e\u57169<\/strong>\u53ef\u4ee5\u770b\u51fa\u666e\u904d\u9810\u6e2c\u7684\u7d50\u679c\u5747\u70ba\u300cLow\u300d\uff0c\u96d6\u7136\u5be6\u969b\u503c\u300cLow\u300d\u4e5f\u662f\u4f54\u5c45\u591a\u6578\uff0c\u4f46\u9019\u908a\u6211\u5011\u53ef\u4ee5\u5408\u7406\u61f7\u7591\u6b64\u8cc7\u6599\u662f\u5426\u767c\u751f\u4e86\u300c\u8cc7\u6599\u4e0d\u5e73\u8861\u300d\u7684\u72c0\u6cc1\uff0c\u5982\u571610<\/strong>\u6240\u793a\u3002<\/p>

\"\"
\u571610. \u8a13\u7df4\u8cc7\u6599\u96c6\u7684\u8a62\u554f\u5ea6\u5206\u5e03\u72c0\u6cc1<\/figcaption><\/figure>
<\/figcaption><\/figure>

\u5f9e\u571610<\/strong>\u767c\u73fe\u6574\u9ad4\u8cc7\u6599\u65bc\u8d77\u59cb\u72c0\u614b\u4fbf\u5177\u6709\u5206\u5e03\u4e0d\u5e73\u5747\u7684\u60c5\u5f62\u7522\u751f\uff0c\u56e0\u6b64\u6211\u5011\u5c07\u5728\u4e0b\u4e00\u7bc7\u7cfb\u52173_\u63d0\u5347\u6a5f\u5668\u5b78\u7fd2\u6e96\u78ba\u5ea6\u7684\u65b9\u6cd5<\/em><\/strong>\u4e2d\uff0c\u4f7f\u7528\u8655\u7406\u8cc7\u6599\u4e0d\u5e73\u8861\u7684\u65b9\u6cd5\u4e26\u9032\u4e00\u6b65\u4ecb\u7d39\u53ca\u5982\u4f55\u4f7f\u7528Log Loss\u6bd4\u8f03\u65b9\u6cd5\u9593\u7684\u6210\u6548\uff0c\u70ba\u5927\u5bb6\u5e36\u4f86\u4e0d\u540c\u7684\u60f3\u6cd5\u4e26\u4f7f\u7528\u6a5f\u5668\u5b78\u7fd2\u63d0\u9ad8\u623f\u4ef2\u696d\u7684\u6f5b\u5728\u6210\u4ea4\u7387!<\/p>

\u4ee5\u4e0a\u5c31\u662f\u7cfb\u52172_\u9032\u968e\u8cc7\u6599\u8655\u7406\u8207\u57fa\u790e\u5efa\u6a21\u7684\u5168\u90e8\u56c9
\u770b\u5b8c\u5f8c\uff0c\u5982\u679c\u89ba\u5f97\u559c\u6b61\uff0c\u4e0d\u59a8\u5e6b\u5fd9\u62cd\u500b\u624b\u00a0!
\u6211\u5011\u4e0b\u6b21\u898b~<\/em><\/strong><\/p><\/blockquote>

\u53c3\u8003\u8cc7\u6599:<\/strong><\/p>

  1. \u7279\u5fb5\u9078\u7528\u53c3\u8003\u8cc7\u6599\uff1a<\/strong>https:\/\/bit.ly\/39ce0jR<\/a><\/li>
  2. Random Forest\u53c3\u8003\u8cc7\u6599\uff1a<\/strong>https:\/\/www.stat.berkeley.edu\/~breiman\/randomforest2001.pdf<\/a><\/li>
  3. XGBoost\u53c3\u8003\u8cc7\u6599\uff1a<\/strong>https:\/\/arxiv.org\/pdf\/1603.02754.pdf<\/a><\/li>
  4. LightGBM\u53c3\u8003\u8cc7\u6599\uff1a<\/strong>https:\/\/papers.nips.cc\/paper\/2017\/file\/6449f44a102fde848669bdd9eb6b76fa-Paper.pdf<\/a><\/li>
  5. \u5982\u4f55\u4f7f\u7528\u6a5f\u5668\u5b78\u7fd2\u63d0\u9ad8\u623f\u4ef2\u696d\u6f5b\u5728\u6210\u4ea4\u7387_\u7cfb\u52171_\u8cc7\u6599\u8655\u7406\u57fa\u672c\u5fc3\u6cd5<\/strong><\/li><\/ol>

    \u4f5c\u8005\uff1a\u9673\u653f\u5ef7\u3001\u738b\u88d5\u840d\u3001\u8b1d\u8c50\u6a8d(\u81fa\u7063\u884c\u92b7\u7814\u7a76\u7279\u9080\u4f5c\u8005)\u3001\u937e\u7693\u8ed2(\u81fa\u7063\u884c\u92b7\u7814\u7a76\u6709\u9650\u516c\u53f8\u5275\u8fa6\u4eba\uff09<\/strong><\/p>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t

    \n\t\t\t\t\t\t
    \n\t\t\t\t\t
    \n\t\t\t
    \n\t\t\t\t\t\t
    \n\t\t\t\t
    \n\t\t\t

    \u66f4\u591a\u5be6\u6230\u6848\u4f8b\u53ca\u60c5\u5883\u597d\u6587\u63a8\u85a6<\/h2>\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
    \n\t\t\t\t
    \n\t\t\t\t\t
    \n\t\t\t\t
    \n\t\t\t
    \n\t\t\t\t
    \n\t\t\t\t

    \n\t\t\t\n\t\t\t\t\u8a02\u623f\u7db2\u4e4b\u71b1\u9580\u98ef\u5e97\u7684\u6f14\u7b97\u908f\u8f2f-\u9ede\u64ca\u7387\u8207\u8a02\u623f\u7d00\u9304\t\t\t<\/a>\n\t\t<\/h3>\n\t\t\t\t
    \n\t\t\t

    \u8a02\u623f\u7db2\u4e4b\u71b1\u9580\u98ef\u5e97\u7684\u6f14\u7b97\u908f\u8f2f-\u9ede\u64ca\u7387\u8207\u8a02\u623f\u7d00\u9304 \u60c5\u5883\uff1a \u7db2\u7f8e\u5c0f\u7f8e3\u500b\u6708\u524d\u5f9e\u6fb3\u6d32\u5e03\u91cc\u65af\u672c\u5ea6\u5047\u56de\u4f86\u3002\u7576\u5730\u5b9c\u4eba\u7684\u6c23\u5019\uff0c\u7368\u7279\u7684\u6587\u5316\u4ee4\u5c0f\u7f8e\u7559\u6200\u5fd8\u8fd4\uff0c\u4f46\u5c0f<\/p>\n\t\t<\/div>\n\t\t\n\t\t\n\t\t\t\u95b1\u8b80\u66f4\u591a \u00bb\t\t<\/a>\n\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/article>\n\t\t\t\t

    \n\t\t\t
    \n\t\t\t\t
    \n\t\t\t\t

    \n\t\t\t\n\t\t\t\t\u5982\u4f55\u6210\u70ba\u8cc7\u6599\u79d1\u5b78\u5bb6!? \u7528 “\u8cc7\u6599\u5206\u6790” \u7684\u65b9\u6cd5\u4f86\u63a2\u8a0e\u200a-\u200a\u7cfb\u52171\uff08\u9644Python\u7a0b\u5f0f\u78bc\uff09\t\t\t<\/a>\n\t\t<\/h3>\n\t\t\t\t
    \n\t\t\t

    \u5982\u4f55\u6210\u70ba\u8cc7\u6599\u79d1\u5b78\u5bb6!? \u7528 “\u8cc7\u6599\u5206\u6790” \u7684\u65b9\u6cd5\u4f86\u63a2\u8a0e\u200a-\u200a\u7cfb\u52171\uff08\u9644Python\u7a0b\u5f0f\u78bc\uff09 \u76f8\u4fe1\u5927\u5bb6\u90fd\u6709\u8033\u805e\u904e\u300c21<\/p>\n\t\t<\/div>\n\t\t\n\t\t\n\t\t\t\u95b1\u8b80\u66f4\u591a \u00bb\t\t<\/a>\n\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/article>\n\t\t\t\t

    \n\t\t\t
    \n\t\t\t\t
    \n\t\t\t\t

    \n\t\t\t\n\t\t\t\t\u5982\u4f55\u8a55\u4f30\u4e0d\u540c\u7522\u54c1\u9069\u5408\u7684\u5ee3\u544a\u65b9\u5f0f?! \u8cc7\u6599\u524d\u8655\u7406_\u96d9\u6a23\u672ct\u6aa2\u5b9a\u6280\u6cd5(\u9644Python\u7a0b\u5f0f\u78bc)\t\t\t<\/a>\n\t\t<\/h3>\n\t\t\t\t
    \n\t\t\t

    \u5982\u4f55\u8a55\u4f30\u4e0d\u540c\u7522\u54c1\u9069\u5408\u7684\u5ee3\u544a\u65b9\u5f0f?! \u8cc7\u6599\u524d\u8655\u7406_\u96d9\u6a23\u672ct\u6aa2\u5b9a\u6280\u6cd5(\u9644Python\u7a0b\u5f0f\u78bc) \u6b64\u7bc7\u5c07\u8b1b\u8ff0\u4f55\u8b02\u96d9\u6a23\u672ct\u6aa2\u5b9a\uff0c\u4e26\u4f7f\u7528\u8a72\u65b9\u6cd5\u5224\u65b7\u4e0d\u540c\u7522<\/p>\n\t\t<\/div>\n\t\t\n\t\t\n\t\t\t\u95b1\u8b80\u66f4\u591a \u00bb\t\t<\/a>\n\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/article>\n\t\t\t\t<\/div>\n\t\t\n\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t<\/div>\n\t\t","protected":false},"excerpt":{"rendered":"

    \u5982\u4f55\u4f7f\u7528\u6a5f\u5668\u5b78\u7fd2\u63d0\u9ad8\u623f\u4ef2\u696d\u6f5b\u5728\u6210\u4ea4\u7387\uff1f\u9032\u968e\u8cc7\u6599\u8655\u7406\u9762\u8207\u57fa\u790e\u5efa\u6a21(\u9644Python\u7a0b\u5f0f\u78bc) \u5148\u56de\u9867\u4e0a\u4e00\u7bc7\u6587\u7ae0\u5427\uff5e\u5982 …<\/p>\n

    \u5982\u4f55\u8a55\u4f30\u4e0d\u540c\u7522\u54c1\u9069\u5408\u7684\u5ee3\u544a\u65b9\u5f0f?! \u8cc7\u6599\u524d\u8655\u7406_\u96d9\u6a23\u672ct\u6aa2\u5b9a\u6280\u6cd5(\u9644Python\u7a0b\u5f0f\u78bc)<\/span> \u7e7c\u7e8c\u95b1\u8b80 \u00bb<\/a><\/p>\n","protected":false},"author":58,"featured_media":0,"comment_status":"closed","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[14],"tags":[],"_links":{"self":[{"href":"https:\/\/tmrmds.co\/wp-json\/wp\/v2\/posts\/25784"}],"collection":[{"href":"https:\/\/tmrmds.co\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/tmrmds.co\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/tmrmds.co\/wp-json\/wp\/v2\/users\/58"}],"replies":[{"embeddable":true,"href":"https:\/\/tmrmds.co\/wp-json\/wp\/v2\/comments?post=25784"}],"version-history":[{"count":8,"href":"https:\/\/tmrmds.co\/wp-json\/wp\/v2\/posts\/25784\/revisions"}],"predecessor-version":[{"id":26476,"href":"https:\/\/tmrmds.co\/wp-json\/wp\/v2\/posts\/25784\/revisions\/26476"}],"wp:attachment":[{"href":"https:\/\/tmrmds.co\/wp-json\/wp\/v2\/media?parent=25784"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/tmrmds.co\/wp-json\/wp\/v2\/categories?post=25784"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/tmrmds.co\/wp-json\/wp\/v2\/tags?post=25784"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}