{"id":98993,"date":"2022-03-15T23:17:32","date_gmt":"2022-03-15T15:17:32","guid":{"rendered":"http:\/\/www.zhushiyao.com\/?p=98993"},"modified":"2022-03-15T23:18:01","modified_gmt":"2022-03-15T15:18:01","slug":"16491bc223","status":"publish","type":"post","link":"http:\/\/www.zhushiyao.com\/?p=98993","title":{"rendered":"\u4e00\u6587\u8bfb\u61c2\u5f02\u5e38\u68c0\u6d4b LOF \u7b97\u6cd5\uff08Python\u4ee3\u7801\uff09"},"content":{"rendered":"
\n
\n

\u672c\u7bc7\u548c\u5927\u5bb6\u4ecb\u7ecd\u4e00\u4e2a\u7ecf\u5178\u7684\u5f02\u5e38\u68c0\u6d4b\u7b97\u6cd5\uff1a\u5c40\u90e8\u79bb\u7fa4\u56e0\u5b50(Local Outlier Factor)\uff0c\u7b80\u79f0LOF\u7b97\u6cd5<\/strong>\u3002<\/p>\n

\u80cc\u666f<\/span><\/h2>\n

Local Outlier Factor\uff08LOF\uff09\u662f\u57fa\u4e8e\u5bc6\u5ea6\u7684\u7ecf\u5178\u7b97\u6cd5\uff08Breuning et. al. 2000\uff09, \u6587\u7ae0\u53d1\u8868\u4e8e SIGMOD 2000, \u5230\u76ee\u524d\u5df2\u7ecf\u6709 3000+ \u7684\u5f15\u7528\u3002<\/p>\n

\u5728 LOF \u4e4b\u524d\u7684\u5f02\u5e38\u68c0\u6d4b\u7b97\u6cd5\u5927\u591a\u662f\u57fa\u4e8e\u7edf\u8ba1\u65b9\u6cd5\u7684\uff0c\u6216\u8005\u662f\u501f\u7528\u4e86\u4e00\u4e9b\u805a\u7c7b\u7b97\u6cd5\u7528\u4e8e\u5f02\u5e38\u70b9\u7684\u8bc6\u522b\uff08\u6bd4\u5982 \uff0cDBSCAN\uff0cOPTICS\uff09\u3002\u8fd9\u4e9b\u65b9\u6cd5\u90fd\u6709\u4e00\u4e9b\u4e0d\u5b8c\u7f8e\u7684\u5730\u65b9\uff1a<\/p>\n

    \n
  • \n
    \u57fa\u4e8e\u7edf\u8ba1\u7684\u65b9\u6cd5\uff1a\u901a\u5e38\u9700\u8981\u5047\u8bbe\u6570\u636e\u670d\u4ece\u7279\u5b9a\u7684\u6982\u7387\u5206\u5e03\uff0c\u8fd9\u4e2a\u5047\u8bbe\u5f80\u5f80\u662f\u4e0d\u6210\u7acb\u7684\u3002<\/section>\n<\/li>\n
  • \n
    \u805a\u7c7b\u65b9\u6cd5\uff1a\u901a\u5e38\u53ea\u80fd\u7ed9\u51fa 0\/1 \u7684\u5224\u65ad\uff08\u5373\uff1a\u662f\u4e0d\u662f\u5f02\u5e38\u70b9\uff09\uff0c\u4e0d\u80fd\u91cf\u5316\u6bcf\u4e2a\u6570\u636e\u70b9\u7684\u5f02\u5e38\u7a0b\u5ea6\u3002<\/section>\n<\/li>\n<\/ul>\n

    \u76f8\u6bd4\u8f83\u800c\u8a00\uff0c\u57fa\u4e8e\u5bc6\u5ea6\u7684LOF\u7b97\u6cd5\u8981\u66f4\u7b80\u5355\u3001\u76f4\u89c2\u3002\u5b83\u4e0d\u9700\u8981\u5bf9\u6570\u636e\u7684\u5206\u5e03\u505a\u592a\u591a\u8981\u6c42\uff0c\u8fd8\u80fd\u91cf\u5316\u6bcf\u4e2a\u6570\u636e\u70b9\u7684\u5f02\u5e38\u7a0b\u5ea6\uff08outlierness\uff09\u3002<\/strong><\/p>\n

    \u4e0b\u9762\u5f00\u59cb\u6b63\u5f0f\u4ecb\u7ecdLOF\u7b97\u6cd5\u3002<\/p>\n

    LOF \u7b97\u6cd5<\/span><\/h2>\n

    \u9996\u5148\uff0c\u57fa\u4e8e\u5bc6\u5ea6\u7684\u79bb\u7fa4\u70b9\u68c0\u6d4b\u65b9\u6cd5\u6709\u4e00\u4e2a\u57fa\u672c\u5047\u8bbe\uff1a\u975e\u79bb\u7fa4\u70b9\u5bf9\u8c61\u5468\u56f4\u7684\u5bc6\u5ea6\u4e0e\u5176\u90bb\u57df\u5468\u56f4\u7684\u5bc6\u5ea6\u7c7b\u4f3c\uff0c\u800c\u79bb\u7fa4\u70b9\u5bf9\u8c61\u5468\u56f4\u7684\u5bc6\u5ea6\u663e\u8457\u4e0d\u540c\u4e8e\u5176\u90bb\u57df\u5468\u56f4\u7684\u5bc6\u5ea6\u3002<\/strong><\/p>\n

    \u4ec0\u4e48\u610f\u601d\u5462\uff1f\u770b\u4e0b\u9762\u56fe\u7247\u611f\u53d7\u4e0b\u3002<\/p>\n

    \"\u4e00\u6587\u8bfb\u61c2\u5f02\u5e38\u68c0\u6d4b<\/figure>\n

    \u96c6\u7fa4 C1<\/code> \u5305\u542b\u4e86 400 \u591a\u4e2a\u70b9\uff0c\u96c6\u7fa4 C2<\/code> \u5305\u542b 100 \u4e2a\u70b9\u3002C1<\/code> \u548c C2<\/code> \u90fd\u662f\u4e00\u7c7b\u96c6\u7fa4\u70b9\uff0c\u533a\u522b\u662f C1<\/code> \u4f4d\u7f6e\u6bd4\u8f83\u96c6\u4e2d\uff0c\u6216\u8005\u8bf4\u5bc6\u5ea6\u6bd4\u8f83\u5927\u3002\u800c\u50cf o1<\/code>\u3001o2<\/code>\u70b9\u5747\u4e3a\u5f02\u5e38\u70b9\uff0c\u56e0\u4e3a\u57fa\u4e8e\u6211\u4eec\u7684\u5047\u8bbe\uff0c\u8fd9\u4e24\u4e2a\u70b9\u5468\u56f4\u7684\u5bc6\u5ea6\u663e\u8457\u4e0d\u540c\u4e8e\u5468\u56f4\u70b9\u7684\u5bc6\u5ea6\u3002<\/p>\n

    LOF \u5c31\u662f\u57fa\u4e8e\u5bc6\u5ea6\u6765\u5224\u65ad\u5f02\u5e38\u70b9\u7684\uff0c\u901a\u8fc7\u7ed9\u6bcf\u4e2a\u6570\u636e\u70b9\u90fd\u5206\u914d\u4e00\u4e2a\u4f9d\u8d56\u4e8e\u90bb\u57df\u5bc6\u5ea6\u7684\u79bb\u7fa4\u56e0\u5b50 LOF\uff0c\u8fdb\u800c\u5224\u65ad\u8be5\u6570\u636e\u70b9\u662f\u5426\u4e3a\u79bb\u7fa4\u70b9\u3002<\/strong> \u5982\u679c \uff0c\u5219\u8be5\u70b9\u4e3a\u79bb\u7fa4\u70b9\uff0c\u5982\u679c \uff0c\u5219\u8be5\u70b9\u4e3a\u6b63\u5e38\u6570\u636e\u70b9\u3002<\/p>\n

    \u90a3\u4ec0\u4e48\u662fLOF\u5462\uff1f<\/p>\n

    \u4e86\u89e3LOF\u524d\uff0c\u5fc5\u987b\u5148\u77e5\u9053\u4e00\u4e0b\u51e0\u4e2a\u57fa\u672c\u6982\u5ff5\uff0c\u56e0\u4e3aLOF\u662f\u57fa\u4e8e\u8fd9\u51e0\u4e2a\u6982\u5ff5\u800c\u6765\u7684\u3002<\/p>\n

    1. k\u90bb\u8fd1\u8ddd\u79bb<\/strong><\/p>\n

    \u5728\u8ddd\u79bb\u6570\u636e\u70b9 \u6700\u8fd1\u7684\u51e0\u4e2a\u70b9\u4e2d\uff0c\u7b2c \u4e2a\u6700\u8fd1\u7684\u70b9\u8ddf\u70b9 \u4e4b\u95f4\u7684\u8ddd\u79bb\u79f0\u4e3a\u70b9 \u7684 K-\u90bb\u8fd1\u8ddd\u79bb\uff0c\u8bb0\u4e3a k-distance (p)\uff0c\u516c\u5f0f\u5982\u4e0b\uff1a<\/p>\n

    <\/section>\n
    \n

    \u70b9 \u4e3a\u8ddd\u79bb\u70b9 \u6700\u8fd1\u7684\u7b2c \u4e2a\u70b9\u3002<\/p>\n<\/blockquote>\n

    \"\u4e00\u6587\u8bfb\u61c2\u5f02\u5e38\u68c0\u6d4b<\/figure>\n

    \u6bd4\u5982\u4e0a\u56fe\u4e2d\uff0c\u8ddd\u79bb\u70b9 \u6700\u8fd1\u7684\u7b2c \u4e2a\u70b9\u662f\u70b9 \u3002<\/p>\n

    \u8fd9\u91cc\u7684\u8ddd\u79bb\u8ba1\u7b97\u53ef\u4ee5\u91c7\u7528\u6b27\u5f0f\u8ddd\u79bb\u3001\u6c49\u660e\u8ddd\u79bb\u3001\u9a6c\u6c0f\u8ddd\u79bb\u7b49\u7b49\u3002\u6bd4\u5982\u7528\u6b27\u5f0f\u8ddd\u79bb\u7684\u8ba1\u7b97\u516c\u5f0f\u5982\u4e0b\uff1a<\/p>\n

    <\/embed><\/section>\n
    \n

    \u8fd9\u91cc\u7684\u91cd\u70b9\u662f\u627e\u5230\u7b2c \u4e2a\u6700\u8fd1\u7684\u90a3\u4e2a\u70b9\uff0c\u7136\u540e\u5e26\u516c\u5f0f\u8ba1\u7b97\u8ddd\u79bb\u3002<\/p>\n<\/blockquote>\n

    2. k\u8ddd\u79bb\u9886\u57df<\/strong><\/p>\n

    \u4ee5\u70b9 \u4e3a\u5706\u5fc3\uff0c\u4ee5k\u90bb\u8fd1\u8ddd\u79bb \u4e3a\u534a\u5f84\u753b\u5706\uff0c\u8fd9\u4e2a\u5706\u4ee5\u5185\u7684\u8303\u56f4\u5c31\u662fk\u8ddd\u79bb\u9886\u57df\uff0c\u516c\u5f0f\u5982\u4e0b\uff1a<\/p>\n

    <\/embed><\/section>\n

    \u8fd8\u662f\u4e0a\u56fe\u6240\u793a\uff0c\u5047\u8bbek=4\uff0c\u90a3\u4e48\u70b9 1-6 \u5747\u662f\u90bb\u57df\u8303\u56f4\u5185\u7684\u70b9\u3002<\/p>\n

    3. \u53ef\u8fbe\u8ddd\u79bb<\/strong><\/p>\n

    \u8fd9\u4e2a\u53ef\u8fbe\u8ddd\u79bb\u5927\u5bb6\u9700\u8981\u7559\u610f\u70b9\uff0c\u70b9 \u5230\u70b9 \u7684\u7b2c \u53ef\u8fbe\u8ddd\u79bb\uff1a<\/p>\n

    <\/embed><\/section>\n

    \u8fd9\u91cc\u8ba1\u7b97 \u5230\u70b9 \u7684\u7b2c \u53ef\u8fbe\u8ddd\u79bb\uff0c\u4f46\u662f\u8981\u4ee5\u70b9 \u4e3a\u4e2d\u5fc3\uff0c\u53d6\u4e00\u4e2a\u6700\u5927\u503c\uff0c\u4e5f\u5c31\u662f\u5728\u70b9 \u4e0e \u7684\u8ddd\u79bb\u3001\u8ddd\u79bb\u70b9 \u6700\u8fd1\u7684\u7b2c \u4e2a\u70b9\u8ddd\u79bb\u4e2d\u53d6\u8f83\u5927\u7684\u4e00\u4e2a\uff0c\u5982\u56fe\u4e0b\u6240\u793a\u3002<\/p>\n

    \"\u4e00\u6587\u8bfb\u61c2\u5f02\u5e38\u68c0\u6d4b<\/figure>\n

    \u8ddd\u79bb \u8fdc\uff0c\u90a3\u4e48\u4e24\u8005\u4e4b\u95f4\u7684\u53ef\u8fbe\u8ddd\u79bb\u5c31\u662f\u5b83\u4eec\u7684\u5b9e\u9645\u8ddd\u79bb\u3002\u5982\u679c\u8ddd\u79bb\u8db3\u591f\u8fd1\uff0c\u5982\u70b9 \uff0c\u5b9e\u9645\u8ddd\u79bb\u5c06\u88ab \u7684 \u8ddd\u79bb\u4ee3\u66ff\u3002\u6240\u6709 \u63a5\u8fd1 \u7684\u7edf\u8ba1\u6ce2\u52a8 \u53ef\u4ee5\u663e\u8457\u51cf\u5c11\uff0c\u8fd9\u53ef\u4ee5\u901a\u8fc7\u53c2\u6570 \u6765\u63a7\u5236\uff0c \u503c\u8d8a\u9ad8\uff0c\u540c\u4e00\u90bb\u57df\u5185\u7684\u70b9\u7684\u53ef\u8fbe\u8ddd\u79bb\u8d8a\u76f8\u4f3c\u3002<\/p>\n

    4. \u5c40\u90e8\u53ef\u8fbe\u5bc6\u5ea6<\/strong><\/p>\n

    \u5148\u7ed9\u51fa\u516c\u5f0f\u3002<\/p>\n

    <\/embed><\/section>\n

    \u6570\u636e\u70b9 \u7684\u5c40\u90e8\u53ef\u8fbe\u5bc6\u5ea6\u5c31\u662f\u57fa\u4e8e \u7684\u6700\u8fd1\u90bb\u7684\u5e73\u5747\u53ef\u8fbe\u8ddd\u79bb\u7684\u5012\u6570\u3002\u8ddd\u79bb\u8d8a\u5927\uff0c\u5bc6\u5ea6\u8d8a\u5c0f\u3002<\/p>\n

    5. \u5c40\u90e8\u5f02\u5e38\u56e0\u5b50<\/strong><\/p>\n

    \u6839\u636e\u5c40\u90e8\u53ef\u8fbe\u5bc6\u5ea6\u7684\u5b9a\u4e49\uff0c\u5982\u679c\u4e00\u4e2a\u6570\u636e\u70b9\u8ddf\u5176\u4ed6\u70b9\u6bd4\u8f83\u758f\u8fdc\u7684\u8bdd\uff0c\u90a3\u4e48\u663e\u7136\u5b83\u7684\u5c40\u90e8\u53ef\u8fbe\u5bc6\u5ea6\u5c31\u5c0f\u3002\u4f46LOF\u7b97\u6cd5\u8861\u91cf\u4e00\u4e2a\u6570\u636e\u70b9\u7684\u5f02\u5e38\u7a0b\u5ea6\uff0c\u5e76\u4e0d\u662f\u770b\u5b83\u7684\u7edd\u5bf9\u5c40\u90e8\u5bc6\u5ea6\uff0c\u800c\u662f\u770b\u5b83\u8ddf\u5468\u56f4\u90bb\u8fd1\u7684\u6570\u636e\u70b9\u7684\u76f8\u5bf9\u5bc6\u5ea6\u3002<\/strong><\/p>\n

    \u8fd9\u6837\u505a\u7684\u597d\u5904\u662f\u53ef\u4ee5\u5141\u8bb8\u6570\u636e\u5206\u5e03\u4e0d\u5747\u5300\u3001\u5bc6\u5ea6\u4e0d\u540c\u7684\u60c5\u51b5\u3002\u5c40\u90e8\u5f02\u5e38\u56e0\u5b50\u5373\u662f\u7528\u5c40\u90e8\u76f8\u5bf9\u5bc6\u5ea6\u6765\u5b9a\u4e49\u7684\u3002\u6570\u636e\u70b9 \u7684\u5c40\u90e8\u76f8\u5bf9\u5bc6\u5ea6\uff08\u5c40\u90e8\u5f02\u5e38\u56e0\u5b50\uff09\u4e3a\u70b9 \u90bb\u57df\u5185\u70b9\u7684\u5e73\u5747\u5c40\u90e8\u53ef\u8fbe\u5bc6\u5ea6\u8ddf\u6570\u636e\u70b9 \u7684\u5c40\u90e8\u53ef\u8fbe\u5bc6\u5ea6\u7684\u6bd4\u503c\uff0c\u5373\uff1a<\/p>\n

    <\/embed><\/section>\n

    LOF\u7b97\u6cd5\u6d41\u7a0b<\/span><\/h2>\n

    \u4e86\u89e3\u4e86 LOF \u7684\u5b9a\u4e49\u4ee5\u540e\uff0c\u6574\u4e2a\u7b97\u6cd5\u4e5f\u5c31\u663e\u800c\u6613\u89c1\u4e86\uff1a<\/p>\n

      \n
    1. \n
      \n

      \u5bf9\u4e8e\u6bcf\u4e2a\u6570\u636e\u70b9\uff0c\u8ba1\u7b97\u5b83\u4e0e\u5176\u5b83\u6240\u6709\u70b9\u7684\u8ddd\u79bb\uff0c\u5e76\u6309\u4ece\u8fd1\u5230\u8fdc\u6392\u5e8f\uff1b<\/p>\n<\/section>\n<\/li>\n

    2. \n
      \n

      \u5bf9\u4e8e\u6bcf\u4e2a\u6570\u636e\u70b9\uff0c\u627e\u5230\u5b83\u7684 k-nearest-neighbor\uff0c\u8ba1\u7b97 LOF \u5f97\u5206;<\/p>\n<\/section>\n<\/li>\n

    3. \n
      \n

      \u5982\u679cLOF\u503c\u8d8a\u5927\uff0c\u8bf4\u660e\u8d8a\u5f02\u5e38\uff0c\u53cd\u4e4b\u5982\u679c\u8d8a\u5c0f\uff0c\u8bf4\u660e\u8d8a\u8d8b\u4e8e\u6b63\u5e38\u3002<\/p>\n<\/section>\n<\/li>\n<\/ol>\n

      \"\u4e00\u6587\u8bfb\u61c2\u5f02\u5e38\u68c0\u6d4b<\/figure>\n

      LOF\u4f18\u7f3a\u70b9<\/span><\/h2>\n

      \u4f18\u70b9<\/strong><\/p>\n

      LOF \u7684\u4e00\u4e2a\u4f18\u70b9\u662f\u5b83\u540c\u65f6\u8003\u8651\u4e86\u6570\u636e\u96c6\u7684\u5c40\u90e8\u548c\u5168\u5c40\u5c5e\u6027\u3002\u5f02\u5e38\u503c\u4e0d\u662f\u6309\u7edd\u5bf9\u503c\u786e\u5b9a\u7684\uff0c\u800c\u662f\u76f8\u5bf9\u4e8e\u5b83\u4eec\u7684\u90bb\u57df\u70b9\u5bc6\u5ea6\u786e\u5b9a\u7684\u3002\u5f53\u6570\u636e\u96c6\u4e2d\u5b58\u5728\u4e0d\u540c\u5bc6\u5ea6\u7684\u4e0d\u540c\u96c6\u7fa4\u65f6\uff0cLOF\u8868\u73b0\u826f\u597d\uff0c\u6bd4\u8f83\u9002\u7528\u4e8e\u4e2d\u7b49\u9ad8\u7ef4\u7684\u6570\u636e\u96c6\u3002<\/p>\n

      \u7f3a\u70b9<\/strong><\/p>\n

      LOF\u7b97\u6cd5\u4e2d\u5173\u4e8e\u5c40\u90e8\u53ef\u8fbe\u5bc6\u5ea6\u7684\u5b9a\u4e49\u5176\u5b9e\u6697\u542b\u4e86\u4e00\u4e2a\u5047\u8bbe\uff0c\u5373\uff1a\u4e0d\u5b58\u5728\u5927\u4e8e\u7b49\u4e8e k \u4e2a\u91cd\u590d\u7684\u70b9\u3002<\/strong><\/p>\n

      \u5f53\u8fd9\u6837\u7684\u91cd\u590d\u70b9\u5b58\u5728\u7684\u65f6\u5019\uff0c\u8fd9\u4e9b\u70b9\u7684\u5e73\u5747\u53ef\u8fbe\u8ddd\u79bb\u4e3a\u96f6\uff0c\u5c40\u90e8\u53ef\u8fbe\u5bc6\u5ea6\u5c31\u53d8\u4e3a\u65e0\u7a77\u5927\uff0c\u4f1a\u7ed9\u8ba1\u7b97\u5e26\u6765\u4e00\u4e9b\u9ebb\u70e6\u3002\u5728\u5b9e\u9645\u5e94\u7528\u65f6\uff0c\u4e3a\u4e86\u907f\u514d\u8fd9\u6837\u7684\u60c5\u51b5\u51fa\u73b0\uff0c\u53ef\u4ee5\u628a k-distance \u6539\u4e3a k-distinct-distance\uff0c\u4e0d\u8003\u8651\u91cd\u590d\u7684\u60c5\u51b5\u3002\u6216\u8005\uff0c\u8fd8\u53ef\u4ee5\u8003\u8651\u7ed9\u53ef\u8fbe\u8ddd\u79bb\u90fd\u52a0\u4e00\u4e2a\u5f88\u5c0f\u7684\u503c\uff0c\u907f\u514d\u53ef\u8fbe\u8ddd\u79bb\u7b49\u4e8e\u96f6\u3002<\/p>\n

      \u53e6\u5916\uff0cLOF \u7b97\u6cd5\u9700\u8981\u8ba1\u7b97\u6570\u636e\u70b9\u4e24\u4e24\u4e4b\u95f4\u7684\u8ddd\u79bb\uff0c\u9020\u6210\u6574\u4e2a\u7b97\u6cd5\u65f6\u95f4\u590d\u6742\u5ea6\u4e3a \u3002\u4e3a\u4e86\u63d0\u9ad8\u7b97\u6cd5\u6548\u7387\uff0c\u540e\u7eed\u6709\u7b97\u6cd5\u5c1d\u8bd5\u6539\u8fdb\u3002FastLOF \uff08Goldstein\uff0c2012\uff09\u5148\u5c06\u6574\u4e2a\u6570\u636e\u968f\u673a\u7684\u5206\u6210\u591a\u4e2a\u5b50\u96c6\uff0c\u7136\u540e\u5728\u6bcf\u4e2a\u5b50\u96c6\u91cc\u8ba1\u7b97 LOF \u503c\u3002\u5bf9\u4e8e\u90a3\u4e9b LOF \u5f02\u5e38\u5f97\u5206\u5c0f\u4e8e\u7b49\u4e8e 1 \u7684\uff0c\u4ece\u6570\u636e\u96c6\u91cc\u5254\u9664\uff0c\u5269\u4e0b\u7684\u5728\u4e0b\u4e00\u8f6e\u5bfb\u627e\u66f4\u5408\u9002\u7684 nearest-neighbor\uff0c\u5e76\u66f4\u65b0 LOF \u503c\u3002<\/p>\n

      Python \u5b9e\u73b0 LOF<\/span><\/h2>\n

      \u6709\u4e24\u4e2a\u5e93\u53ef\u4ee5\u8ba1\u7b97LOF\uff0c\u5206\u522b\u662fPyOD<\/code>\u548cSklearn<\/code>\uff0c\u4e0b\u9762\u5206\u522b\u4ecb\u7ecd\u3002<\/p>\n

      \u4f7f\u7528pyod<\/code>\u81ea\u5e26\u7684\u65b9\u6cd5\u751f\u6210200\u4e2a\u8bad\u7ec3\u6837\u672c\u548c100\u4e2a\u6d4b\u8bd5\u6837\u672c\u7684\u6570\u636e\u96c6\u3002\u6b63\u6001\u6837\u672c\u7531\u591a\u5143\u9ad8\u65af\u5206\u5e03\u751f\u6210\uff0c\u5f02\u5e38\u6837\u672c\u662f\u4f7f\u7528\u5747\u5300\u5206\u5e03\u751f\u6210\u7684\u3002<\/p>\n

      \u8bad\u7ec3\u548c\u6d4b\u8bd5\u6570\u636e\u96c6\u90fd\u6709 5 \u4e2a\u7279\u5f81\uff0c10% \u7684\u884c\u88ab\u6807\u8bb0\u4e3a\u5f02\u5e38\u3002\u5e76\u4e14\u5728\u6570\u636e\u4e2d\u6dfb\u52a0\u4e86\u4e00\u4e9b\u968f\u673a\u566a\u58f0\uff0c\u8ba9\u5b8c\u7f8e\u5206\u79bb\u6b63\u5e38\u70b9\u548c\u5f02\u5e38\u70b9\u53d8\u5f97\u7a0d\u5fae\u56f0\u96be\u4e00\u4e9b\u3002<\/p>\n

      from<\/span>\u00a0pyod.utils.data\u00a0import<\/span>\u00a0generate_data\nimport<\/span>\u00a0numpy\u00a0as<\/span>\u00a0np\nX_train,\u00a0y_train,\u00a0X_test,\u00a0y_test\u00a0=\u00a0\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0generate_data(n_train=200<\/span>,\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0n_test=100<\/span>,\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0n_features=5<\/span>,\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0contamination=0.1<\/span>,\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0random_state=3<\/span>)\u00a0\nX_train\u00a0=\u00a0X_train\u00a0*\u00a0np.random.uniform(0<\/span>,\u00a01<\/span>,\u00a0size=X_train.shape)\nX_test\u00a0=\u00a0X_test\u00a0*\u00a0np.random.uniform(0<\/span>,1<\/span>,\u00a0size=X_test.shape)\n<\/code><\/pre>\n
      PyOD<\/strong><\/p>\n
      \u4e0b\u9762\u5c06\u8bad\u7ec3\u6570\u636e\u62df\u5408\u4e86 LOF \u6a21\u578b\u5e76\u5c06\u5176\u5e94\u7528\u4e8e\u5408\u6210\u6d4b\u8bd5\u6570\u636e\u3002<\/p>\n
      \u5728 PyOD<\/code> \u4e2d\uff0c\u6709\u4e24\u4e2a\u5173\u952e\u65b9\u6cd5\uff1adecision_function<\/code> \u548c predict<\/code>\u3002<\/p>\n