{"id":114974,"date":"2019-05-22T11:07:06","date_gmt":"2019-05-22T03:07:06","guid":{"rendered":"http:\/\/www.fagao.me\/?p=114974"},"modified":"2019-05-22T11:07:06","modified_gmt":"2019-05-22T03:07:06","slug":"%e8%be%be%e8%a7%82%e6%95%b0%e6%8d%ae-%e6%90%9c%e4%bd%a0%e6%89%80%e6%83%b3-%e4%b9%8b%e7%94%a8%e6%88%b7%e6%90%9c%e7%b4%a2%e6%84%8f%e5%9b%be%e8%af%86%e5%88%ab","status":"publish","type":"post","link":"https:\/\/www.fagao.me\/p\/114974.htm","title":{"rendered":"\u8fbe\u89c2\u6570\u636e \u641c\u4f60\u6240\u60f3 \u4e4b\u7528\u6237\u641c\u7d22\u610f\u56fe\u8bc6\u522b"},"content":{"rendered":"<p><img decoding=\"async\" src=\"https:\/\/www.fagao.me\/images\/uploads\/2017_001.jpg\" alt=\"\"><\/p>\n<p>\u4eba\u7c7b\u81ea\u8bde\u751f\u4ee5\u6765\u5c31\u4f34\u968f\u7740\u5404\u79cd\u4fe1\u606f\u7684\u751f\u4ea7\u548c\u83b7\u53d6\uff0c\u5982\u4eca\u8fd9\u4e2a\u4fe1\u606f\u7206\u70b8\u7684 DT \u65f6\u4ee3\uff0c\u4eba\u4eec\u66f4\u662f\u88ab\u5404\u79cd\u4fe1\u606f\u6240\u5305\u56f4\u3002\u6211\u4eec\u77e5\u9053\uff0c\u4eba\u83b7\u53d6\u4fe1\u606f\u7684\u65b9\u5f0f\u4e3b\u8981\u6709\u88ab\u52a8\u83b7\u53d6\u548c\u4e3b\u52a8\u83b7\u53d6\u4e24\u79cd\uff0c\u5176\u4e2d\u88ab\u52a8\u83b7\u53d6\u5c31\u662f\u63a8\u8350\u7684\u65b9\u5f0f\u3001\u4e3b\u52a8\u83b7\u53d6\u5c31\u662f\u641c\u7d22\u7684\u65b9\u5f0f\u3002<\/p>\n<p>\u83b7\u53d6\u4fe1\u606f\u662f\u4eba\u7c7b\u8ba4\u77e5\u4e16\u754c\u3001\u751f\u5b58\u53d1\u5c55\u7684\u521a\u9700\uff0c\u641c\u7d22\u5c31\u662f\u6700\u660e\u786e\u7684\u4e00\u79cd\u65b9\u5f0f\uff0c\u5176\u4f53\u73b0\u7684\u52a8\u4f5c\u5c31\u662f\u201c\u51fa\u53bb\u627e\u201d\uff0c\u627e\u98df\u7269\u3001\u627e\u5730\u70b9\u7b49\uff0c\u5230\u4e86\u4e92\u8054\u7f51\u65f6\u4ee3\uff0c<a class=\"tag-links\" href=\"https:\/\/www.fagao.me\/p\/t\/121\">\u641c\u7d22\u5f15\u64ce<\/a>\uff08Search Engine\uff09\u5c31\u662f\u6ee1\u8db3\u627e\u4fe1\u606f\u8fd9\u4e2a\u9700\u6c42\u7684\u6700\u597d\u5de5\u5177\uff0c\u4f60\u8f93\u5165\u60f3\u8981\u627e\u7684\u5185\u5bb9\uff08\u5373\u5728\u641c\u7d22\u6846\u91cc\u8f93\u5165\u67e5\u8be2\u8bcd\uff0c\u6216\u79f0\u4e3a Query\uff09\uff0c<a class=\"tag-links\" href=\"https:\/\/www.fagao.me\/p\/t\/121\">\u641c\u7d22\u5f15\u64ce<\/a>\u5feb\u901f\u7684\u7ed9\u4f60\u6700\u597d\u7684\u7ed3\u679c\uff0c\u8fd9\u6837\u7684\u521a\u9700\u50ac\u751f\u4e86\u8c37\u6b4c\u3001\u767e\u5ea6\u8fd9\u6837\u7684\u4e92\u8054\u7f51\u5de8\u5934\u3002<\/p>\n<p>\u6b64\u6b21\u7ed3\u5408\u8fbe\u89c2\u5728\u5782\u76f4<a class=\"tag-links\" href=\"https:\/\/www.fagao.me\/p\/t\/121\">\u641c\u7d22\u5f15\u64ce<\/a>\u5efa\u8bbe\u65b9\u9762\u7684\u7ecf\u9a8c\uff0c\u56f4\u7ed5\u4ee5\u4e0b\u5185\u5bb9\u8fdb\u884c\u5c55\u5f00\uff1a<\/p>\n<ol>\n<li>\u7528\u6237\u641c\u7d22\u610f\u56fe\u7684\u7406\u89e3\u53ca\u5176\u96be\u70b9\u89e3\u6790<\/li>\n<li>\u5982\u4f55\u8bc6\u522b\u7528\u6237\u641c\u7d22\u610f\u56fe<\/li>\n<li>\u8fbe\u89c2\u6570\u636e\u7528\u6237\u641c\u7d22\u610f\u56fe\u7406\u89e3\u5f15\u64ce\u4ecb\u7ecd<\/li>\n<\/ol>\n<p><span>\u4e00\u3001\u7528\u6237\u641c\u7d22\u610f\u56fe\u7684\u7406\u89e3\u53ca\u5176\u96be\u70b9\u89e3\u6790<\/span><\/p>\n<p>\u9996\u5148\uff0c\u4ece\u504f\u6280\u672f\u7684\u89d2\u5ea6\u6765\u770b\u770b<a class=\"tag-links\" href=\"https:\/\/www.fagao.me\/p\/t\/121\">\u641c\u7d22\u5f15\u64ce<\/a>\u53d1\u5c55\u7684\u51e0\u4e2a\u9636\u6bb5\uff1a<\/p>\n<p><strong>\u7b2c\u4e00\u4e2a\u9636\u6bb5<\/strong>\uff0c\u4f7f\u7528\u5012\u6392\u7d22\u5f15\u89e3\u51b3\u5339\u914d\u7684\u6548\u7387\u95ee\u9898\uff0c\u4f7f\u7528\u6587\u6863\u6a21\u578b\u89e3\u51b3\u57fa\u672c\u7684\u76f8\u5173\u6027\uff0c\u4f7f<a class=\"tag-links\" href=\"https:\/\/www.fagao.me\/p\/t\/121\">\u641c\u7d22\u5f15\u64ce<\/a>\u53d8\u5f97\u53ef\u7528\u3001\u53ef\u6269\u5c55\uff0c\u4ee3\u8868\u6bd4\u5982 Infoseek \u3002\u4f46\u8fd9\u4e00\u9636\u6bb5\u53ea\u4fdd\u8bc1\u4e86\u57fa\u672c\u7684\u6587\u5b57\u76f8\u5173\u6027\uff0c\u641c\u7d22\u7684\u771f\u6b63\u6548\u679c\u662f\u65e0\u6cd5\u4fdd\u8bc1\u7684\u3002<\/p>\n<p><strong>\u7b2c\u4e8c\u4e2a\u9636\u6bb5<\/strong>\uff0c\u4f7f\u7528\u8d85\u94fe\u6a21\u578b\uff0c\u6bd4\u5982\u8c37\u6b4c\u7684 pagerank \u548c\u767e\u5ea6\u7684\u8d85\u94fe\u5206\u6790\u3002\u89e3\u51b3\u6743\u5a01\u6027\u95ee\u9898\uff0c\u4f7f\u641c\u7d22\u8d28\u91cf\u63d0\u5347\u4e00\u4e2a\u6863\u6b21\u3002\u4ece\u8fd9\u4e00\u9636\u6bb5<a class=\"tag-links\" href=\"https:\/\/www.fagao.me\/p\/t\/121\">\u641c\u7d22\u5f15\u64ce<\/a>\u5f00\u59cb\u5feb\u901f\u666e\u53ca\u4e0e\u5e76\u8fdb\u5165\u5546\u4e1a\u5316\uff0c\u4e3a\u8c37\u6b4c\u548c\u767e\u5ea6\u8fd9\u6837\u7684\u516c\u53f8\u5e26\u6765\u4e86\u4e30\u539a\u7684\u5229\u6da6\u3002<\/p>\n<p><strong>\u7b2c\u4e09\u4e2a\u9636\u6bb5<\/strong>\uff0c\u4e00\u65b9\u9762\u4f7f\u7528\u66f4\u590d\u6742\u7684\u89c4\u5219\u548c\u673a\u5668\u5b66\u4e60\u6392\u5e8f\u6a21\u578b\uff0c\u7efc\u5408\u8003\u8651\u4e86\u7528\u6237\u7684\u884c\u4e3a\u7279\u5f81\uff0c\u5982\u5546\u54c1\u8bc4\u8bba\u3001\u70b9\u8d5e\u3001\u6536\u85cf\u3001\u8d2d\u4e70\u7b49\uff0c\u4f7f\u5f97<a class=\"tag-links\" href=\"https:\/\/www.fagao.me\/p\/t\/121\">\u641c\u7d22\u5f15\u64ce<\/a>\u7684\u7ed3\u679c\u518d\u6b21\u63d0\u5347\u4e00\u4e2a\u6863\u6b21\uff0c\u8fd9\u4e9b\u5728\u7535\u5546\u7b49\u5782\u76f4\u641c\u7d22\u4e0a\u8868\u73b0\u7684\u4f1a\u66f4\u52a0\u660e\u663e\uff1b\u53e6\u4e00\u65b9\u9762\uff0c\u57fa\u4e8e\u5404\u79cd\u5148\u8fdb\u7684\u81ea\u7136\u8bed\u8a00\u5904\u7406\u6280\u672f\uff0c\u5145\u5206\u6316\u6398\u7528\u6237\u641c\u7d22\u884c\u4e3a\u65e5\u5fd7\uff0c\u5bf9 query \u8fdb\u884c\u5206\u6790\u6539\u5199\u4ee5\u53ec\u56de\u66f4\u591a\u66f4\u597d\u7684\u7ed3\u679c\u3002<\/p>\n<p><strong>\u7b2c\u56db\u4e2a\u9636\u6bb5<\/strong>\uff0c\u4ece\u201c\u6709\u6846\u201d\u641c\u7d22\u65f6\u4ee3\u6b65\u5165\u66f4\u52a0\u4eba\u5de5\u667a\u80fd\u7684\u201c\u65e0\u6846\u201d\u641c\u7d22\u65f6\u4ee3\u3002\u4eba\u673a\u4ea4\u4e92\u65b9\u5f0f\u4e5f\u5c06\u66f4\u591a\u7684\u662f\u95ee\u7b54\u5f0f\u7684\u81ea\u7136\u8bed\u8a00\u52a0\u8bed\u97f3\u7684\u65b9\u5f0f\uff0c\u800c<a class=\"tag-links\" href=\"https:\/\/www.fagao.me\/p\/t\/121\">\u641c\u7d22\u5f15\u64ce<\/a>\u4e5f\u66f4\u50cf\u4e00\u4e2a\u667a\u80fd\u673a\u5668\u4eba\uff0c\u7406\u89e3\u4eba\u7684\u81ea\u7136\u8bed\u8a00\u95ee\u9898\uff0c\u63d0\u4f9b\u66f4\u52a0\u76f4\u63a5\u6709\u6548\u7684\u77e5\u8bc6\u548c\u7b54\u6848\u3002\u8fd9\u4e00\u9636\u6bb5\u76ee\u524d\u5c1a\u5904\u4e8e\u8d77\u6b65\u9636\u6bb5\uff0c\u8c37\u6b4c\u3001Amazon \u4ee5\u53ca\u4e00\u4e9b\u4f18\u79c0\u7684\u521b\u4e1a\u516c\u53f8\u90fd\u5728\u8fdb\u884c\u79ef\u6781\u7684\u63a2\u7d22\u3002<\/p>\n<p><a class=\"tag-links\" href=\"https:\/\/www.fagao.me\/p\/t\/121\">\u641c\u7d22\u5f15\u64ce<\/a>\u6d89\u53ca\u7684\u6280\u672f\u975e\u5e38\u7e41\u590d\uff0c\u65e2\u6709\u5de5\u7a0b\u67b6\u6784\u65b9\u9762\u7684\uff0c\u53c8\u6709\u7b97\u6cd5\u7b56\u7565\u65b9\u9762\u7684\u3002\u7efc\u5408\u6765\u8bb2\uff0c\u4e00\u4e2a<a class=\"tag-links\" href=\"https:\/\/www.fagao.me\/p\/t\/121\">\u641c\u7d22\u5f15\u64ce<\/a>\u7684\u6280\u672f\u6784\u5efa\u4e3b\u8981\u5305\u542b\u4e09\u5927\u90e8\u5206\uff1a<\/p>\n<ol>\n<li>\u5bf9 query \u7684\u7406\u89e3\uff1b<\/li>\n<li>\u5bf9\u5185\u5bb9\uff08\u6587\u6863\uff09\u7684\u7406\u89e3\uff1b<\/li>\n<li>\u5bf9 query \u548c\u5185\u5bb9\uff08\u6587\u6863\uff09\u7684\u5339\u914d\u548c\u6392\u5e8f\u3002<\/li>\n<\/ol>\n<p>\u672c\u7247\u6587\u7ae0\u4e3b\u8981\u63a2\u8ba8\u5176\u4e2d\u7684 Query Understanding\uff0c\u5373\u5bf9 query \u7684\u7406\u89e3\u3002\u5bf9 query \u7684\u7406\u89e3\uff0c \u6362\u53e5\u8bdd\u8bf4\u5c31\u662f\u5bf9\u7528\u6237\u641c\u7d22\u610f\u56fe\u7684\u7406\u89e3\u3002\u5148\u770b\u5782\u76f4\u641c\u7d22\u4e2d\u7684\u4e00\u4e9b\u4f8b\u5b50\uff1a<\/p>\n<p>\u201c\u9644\u8fd1\u7684\u7279\u4ef7\u9152\u5e97\u201d<\/p>\n<p>\u201c\u4e0a\u6d77\u5230\u626c\u5dde\u9ad8\u901f\u600e\u4e48\u8d70\u201d<\/p>\n<p>\u201c\u5c0f\u9f99\u867e\u6700\u65b0\u62a5\u4ef7\u201d<\/p>\n<p>\u201c\u534e\u4e3a\u6700\u65b0\u6b3e\u624b\u673a\u201d<\/p>\n<p>\u201c\u6c34\u201d<\/p>\n<p>\u8fd9\u51e0\u4e2a\u4f8b\u5b50\u90fd\u4e0d\u80fd\u76f4\u63a5\u6839\u636e query \u7684\u5b57\u9762\u610f\u601d\u53bb\u641c\u7d22\uff0c\u800c\u662f\u8981\u7406\u89e3\u7528\u6237\u8f93\u5165\u6587\u5b57\u80cc\u540e\u7684\u771f\u5b9e\u610f\u56fe\u3002\u4e0d\u8fc7\u8981\u51c6\u786e\u7406\u89e3 query \u80cc\u540e\u7684\u7528\u6237\u641c\u7d22\u610f\u56fe\u53ef\u4e0d\u662f\u90a3\u4e48\u5bb9\u6613\u7684\u3002<\/p>\n<p><strong>\u6211\u4eec\u6765\u5206\u6790\u4e00\u4e0b\u7406\u89e3\u7528\u6237\u641c\u7d22\u8bcd\u80cc\u540e\u7684\u771f\u5b9e\u610f\u56fe\u8bc6\u522b\u5b58\u5728\u54ea\u4e9b\u96be\u70b9\uff1a<\/strong><\/p>\n<ol>\n<li>\u7528\u6237\u8f93\u5165\u4e0d\u89c4\u8303\uff0c\u8f93\u5165\u65b9\u5f0f\u591a\u6837\u5316\uff0c\u4f7f\u7528\u81ea\u7136\u8bed\u8a00\u67e5\u8be2\uff0c\u751a\u81f3\u975e\u6807\u51c6\u7684\u81ea\u7136\u8bed\u8a00\u3002\u6bd4\u5982\u4e0a\u9762\u63d0\u5230\u7684\u201c\u9644\u8fd1\u7684\u7279\u4ef7\u9152\u5e97\u201d \u3001\u201c\u4e0a\u6d77\u5230\u626c\u5dde\u9ad8\u901f\u600e\u4e48\u8d70\u201d\u90fd\u662f\u81ea\u7136\u8bed\u8a00\u67e5\u8be2\u7684\u4f8b\u5b50\uff0c\u53c8\u5982 \u201c\u62ab\u661f ( ) \u6708\u201d\u3001\u201c\u543e\u5c1d\u7ec8\u65e5\u800c\u601d\u77e3, \u4e0b\u9762\u201c\u3002<\/li>\n<li>\u7528\u6237\u7684\u67e5\u8be2\u8bcd\u8868\u73b0\u51fa\u591a\u610f\u56fe\uff0c\u6bd4\u5982\u7528\u6237\u641c\u7d22\u201c\u53d8\u5f62\u91d1\u521a\u201d\uff0c\u662f\u6307\u53d8\u5f62\u91d1\u521a\u7684\u7535\u5f71\u8fd8\u662f\u6e38\u620f\uff1f \u641c\u7d22\u201c\u4ed9\u5251\u5947\u4fa0\u4f20\u201d\u662f\u6307\u6e38\u620f\u8fd8\u662f\u6e38\u620f\u8f6f\u4ef6\uff1f \u7535\u5f71\uff1f \u5c0f\u8bf4\uff1f \u7535\u5546\u7f51\u7ad9\u641c\u7d22\u201c\u6c34\u201d\u662f\u6307\u77ff\u6cc9\u6c34\uff1f\u8fd8\u662f\u5973\u751f\u7528\u7684\u62a4\u80a4\u6c34\uff1f<\/li>\n<li>\u610f\u56fe\u5f3a\u5ea6\uff0c\u8868\u73b0\u4e3a\u4e0d\u540c\u7528\u6237\u5bf9\u76f8\u540c\u7684\u67e5\u8be2\u6709\u4e0d\u540c\u7684\u9700\u6c42\u5f3a\u5ea6\u3002\u6bd4\u5982\uff1a\u5bab\u4fdd\u9e21\u4e01\u3002\u5bab\u4fdd\u9e21\u4e01\u83dc\uff0c\u83dc\u8c31\u9700\u6c42\u5360 90%\u3002\u5bab\u4fdd\u9e21\u4e01\u6b4c\u66f2\uff0c\u6b4c\u66f2\u4e0b\u8f7d\u9700\u6c42\u5360 10%\u3002\u53c8\u6bd4\u5982\uff1a\u8377\u5858\u6708\u8272\u3002\u8377\u5858\u6708\u8272\u6b4c\u66f2\uff0c\u6b4c\u66f2\u4e0b\u8f7d\u9700\u6c42\u5360 70%\u3002\u8377\u5858\u6708\u8272\u5c0f\u533a\uff0c\u623f\u4ea7\u9700\u6c42\u5360 20%\u3002\u8377\u5858\u6708\u8272\u83dc\uff0c\u83dc\u8c31\u9700\u6c42\u5360 10%\u3002<\/li>\n<li>\u610f\u56fe\u5b58\u5728\u65f6\u6548\u6027\u53d8\u5316\uff0c\u5c31\u662f\u968f\u7740\u65f6\u95f4\u7684\u63a8\u79fb\u4e00\u4e9b\u67e5\u8be2\u8bcd\u7684\u610f\u56fe\u4f1a\u53d1\u751f\u53d8\u5316\u3002\u6bd4\u5982\uff1a\u534e\u4e3a P10 \u56fd\u884c\u7248 3 \u6708 24 \u65e5\u4e0a\u5e02\u30023 \u6708 21 \u65e5\u7684\u67e5\u8be2\u610f\u56fe\uff1a\u65b0\u95fb 90%\uff0c\u767e\u79d1 10%3 \u6708 24 \u65e5\u7684\u67e5\u8be2\u610f\u56fe\uff1a\u65b0\u95fb 70%\uff0c\u8d2d\u4e70 25%\uff0c\u767e\u79d1 5%5 \u6708 1 \u65e5\u7684\u67e5\u8be2\u610f\u56fe\uff1a\u8d2d\u4e70 50%\uff0c\u8d44\u8baf 40%\uff0c\u5176\u4ed6 10%5 \u5e74\u4ee5\u540e\u7684\u67e5\u8be2\u610f\u56fe\uff1a\u767e\u79d1 100%\u3002<\/li>\n<li>\u6570\u636e\u51b7\u542f\u52a8\u7684\u95ee\u9898\uff0c\u7528\u6237\u884c\u4e3a\u6570\u636e\u8f83\u5c11\u65f6\uff0c\u5f88\u96be\u51c6\u786e\u83b7\u53d6\u7528\u6237\u7684\u641c\u7d22\u610f\u56fe\u3002<\/li>\n<li>\u6ca1\u6709\u56fa\u5b9a\u7684\u8bc4\u4f30\u7684\u6807\u51c6\uff0cCTR\u3001MAP\u3001MRR\u3001nDCG \u8fd9\u4e9b\u53ef\u4ee5\u91cf\u5316\u7684\u6307\u6807\u4e3b\u8981\u662f\u9488\u5bf9<a class=\"tag-links\" href=\"https:\/\/www.fagao.me\/p\/t\/121\">\u641c\u7d22\u5f15\u64ce<\/a>\u7684\u6574\u4f53\u6548\u679c\u7684\uff0c\u5177\u4f53\u5230\u7528\u6237\u610f\u56fe\u7684\u9884\u6d4b\u4e0a\u5e76\u6ca1\u6709\u6807\u51c6\u7684\u6307\u6807\u3002<\/li>\n<\/ol>\n<p><span>\u4e8c\u3001\u5982\u4f55\u8bc6\u522b\u7528\u6237\u641c\u7d22\u610f\u56fe<\/span><\/p>\n<p>\u9996\u5148\u6211\u4eec\u6765\u770b\u4e00\u4e0b\u7528\u6237\u641c\u7d22\u610f\u56fe\u6709\u54ea\u4e9b\u5206\u7c7b\u3002\u4e00\u822c\u628a\u641c\u7d22\u610f\u56fe\u5f52\u7c7b\u4e3a 3 \u79cd\u7c7b\u578b\uff1a\u5bfc\u822a\u7c7b\u3001\u4fe1\u606f\u7c7b\u548c\u4e8b\u52a1\u7c7b\u3002\u96c5\u864e\u7684\u7814\u7a76\u4eba\u5458\u5728\u6b64\u57fa\u7840\u4e0a\u505a\u4e86\u7ec6\u5316\uff0c<strong>\u5c06\u7528\u6237\u641c\u7d22\u610f\u56fe\u5212\u5206\u5982\u4e0b\u7c7b\u522b<\/strong>\uff1a<\/p>\n<p><strong>1. <\/strong><strong>\u5bfc\u822a\u7c7b<\/strong>\uff1a\u7528\u6237\u660e\u786e\u8981\u53bb\u67d0\u4e2a\u7ad9\u70b9\uff0c\u4f46\u53c8\u4e0d\u60f3\u81ea\u5df1\u8f93\u5165 URL\uff0c\u6bd4\u5982\u7528\u6237\u641c\u7d22\u201c\u65b0\u6d6a\u7f51\u201c\u3002<\/p>\n<p><strong>2. <\/strong><strong>\u4fe1\u606f\u7c7b<\/strong>\uff1a\u53ef\u4ee5\u7ec6\u5206\u4e3a\u5982\u4e0b\u51e0\u79cd\u5b50\u7c7b\u578b\u3002<\/p>\n<p><strong>1\uff09\u76f4\u63a5\u578b<\/strong>\uff1a\u7528\u6237\u60f3\u77e5\u9053\u5173\u4e8e\u4e00\u4e2a\u8bdd\u9898\u67d0\u4e2a\u65b9\u9762\u660e\u786e\u7684\u4fe1\u606f\uff0c\u6bd4\u5982\u201c\u5730\u7403\u4e3a\u4ec0\u4e48\u662f\u5706\u7684\u201d\u3001\u201c\u54ea\u4e9b\u6c34\u679c\u7ef4\u751f\u7d20\u542b\u91cf\u9ad8\u201d\u3002<\/p>\n<p><strong>2\uff09\u95f4\u63a5\u578b<\/strong>\uff1a\u7528\u6237\u60f3\u4e86\u89e3\u5173\u4e8e\u67d0\u4e2a\u8bdd\u9898\u7684\u4efb\u610f\u65b9\u9762\u7684\u4fe1\u606f\uff0c\u6bd4\u5982\u7c89\u4e1d\u641c\u7d22\u201c\u9ec4\u6653\u660e\u201d\u3002<\/p>\n<p><strong>3\uff09\u5efa\u8bae\u578b<\/strong>\uff1a\u7528\u6237\u5e0c\u671b\u80fd\u591f\u641c\u7d22\u5230\u4e00\u4e9b\u5efa\u8bae\u3001\u610f\u89c1\u6216\u8005\u67d0\u65b9\u9762\u7684\u6307\u5bfc\uff0c\u6bd4\u5982\u201c\u5982\u4f55\u9009\u80a1\u7968\u201d\u3002<\/p>\n<p><strong>4\uff09\u5b9a\u4f4d\u578b<\/strong>\uff1a\u7528\u6237\u5e0c\u671b\u4e86\u89e3\u5728\u73b0\u5b9e\u751f\u6d3b\u4e2d\u54ea\u91cc\u53ef\u4ee5\u627e\u5230\u67d0\u4e9b\u4ea7\u54c1\u6216\u670d\u52a1\uff0c\u6bd4\u5982\u201c\u6c7d\u8f66\u7ef4\u4fee\u201d\u3002<\/p>\n<p><img decoding=\"async\" src=\"https:\/\/www.fagao.me\/images\/uploads\/F002.jpg\" alt=\"\"><\/p>\n<p><strong>5\uff09\u5217\u8868\u578b<\/strong>\uff1a\u7528\u6237\u5e0c\u671b\u627e\u5230\u4e00\u6279\u80fd\u591f\u6ee1\u8db3\u9700\u6c42\u7684\u4fe1\u606f\uff0c\u6bd4\u5982\u201c\u9646\u5bb6\u5634\u9644\u8fd1\u7684\u9152\u5e97\u201d\u3002<\/p>\n<p><strong>3. <\/strong><strong>\u8d44\u6e90\u7c7b<\/strong>\uff1a\u8fd9\u79cd\u7c7b\u578b\u7684\u641c\u7d22\u76ee\u7684\u662f\u5e0c\u671b\u80fd\u591f\u4ece\u7f51\u4e0a\u83b7\u53d6\u67d0\u79cd\u8d44\u6e90\uff0c\u53c8\u53ef\u4ee5\u7ec6\u5206\u4e3a\u4ee5\u4e0b\u51e0\u79cd\u5b50\u7c7b\u578b\u3002<\/p>\n<p><strong>1\uff09\u4e0b\u8f7d\u578b<\/strong>\uff1a\u5e0c\u671b\u4ece\u7f51\u7edc\u67d0\u4e2a\u5730\u65b9\u4e0b\u8f7d\u60f3\u8981\u7684\u4ea7\u54c1\u6216\u8005\u670d\u52a1\uff0c\u6bd4\u5982\u201cUSB \u9a71\u52a8\u4e0b\u8f7d\u201d\u3002<\/p>\n<p><strong>2\uff09\u5a31\u4e50\u578b<\/strong>\uff1a\u7528\u6237\u51fa\u4e8e\u6d88\u9063\u7684\u76ee\u7684\u5e0c\u671b\u83b7\u5f97\u4e00\u4e9b\u6709\u5173\u4fe1\u606f\uff0c\u6bd4\u5982\u201c\u76ca\u667a\u5c0f\u6e38\u620f\u201d\u3002<\/p>\n<p><strong>3\uff09\u4ea4\u4e92\u578b<\/strong>\uff1a\u7528\u6237\u5e0c\u671b\u4f7f\u7528\u67d0\u4e2a\u8f6f\u4ef6\u6216\u670d\u52a1\u63d0\u4f9b\u7684\u7ed3\u679c\uff0c\u7528\u6237\u5e0c\u671b\u627e\u5230\u4e00\u4e2a\u7f51\u7ad9\uff0c\u8fd9\u4e2a\u7f51\u7ad9\u4e0a\u53ef\u4ee5\u76f4\u63a5\u8ba1\u7b97\u623f\u8d37\u5229\u606f\u3002<\/p>\n<p><strong>4\uff09\u83b7\u53d6\u578b<\/strong>\uff1a\u7528\u6237\u5e0c\u671b\u83b7\u53d6\u4e00\u79cd\u8d44\u6e90\uff0c\u8fd9\u79cd\u8d44\u6e90\u7684\u4f7f\u7528\u573a\u5408\u4e0d\u9650\u4e8e\u7535\u8111\uff0c\u6bd4\u5982\u201c\u9ea6\u5f53\u52b3\u4f18\u60e0\u5238\u201d\uff0c\u7528\u6237\u5e0c\u671b\u641c\u5230\u67d0\u4e2a\u4ea7\u54c1\u7684\u6298\u6263\u5238\uff0c\u6253\u5370\u540e\u5728\u73b0\u5b9e\u751f\u6d3b\u4e2d\u4f7f\u7528\u3002<\/p>\n<p>\u67e5\u8be2\u610f\u56fe\u7406\u89e3\u7684\u8fc7\u7a0b\u5c31\u662f\u7ed3\u5408\u7528\u6237\u5386\u53f2\u884c\u4e3a\u6570\u636e\u5bf9 query \u8fdb\u884c\u5404\u79cd\u5206\u6790\u5904\u7406\u7684\u8fc7\u7a0b\uff0c\u5305\u62ec\u67e5\u8be2\u7ea0\u9519\u3001\u67e5\u8be2\u8bcd\u81ea\u52a8\u63d0\u793a\u3001\u67e5\u8be2\u6269\u5c55\uff0c\u67e5\u8be2\u81ea\u52a8\u5206\u7c7b\u3001\u8bed\u4e49\u6807\u7b7e\u7b49\u3002<\/p>\n<p>\u4e0a\u9762\u8fd9\u5f20\u56fe\u662f\u4e00\u4e2a\u5177\u4f53\u7684\u4f8b\u5b50\u8bf4\u660e query understanding \u7684\u5de5\u4f5c\u8fc7\u7a0b\uff1a<\/p>\n<p>\u7a0d\u5fae\u89e3\u91ca\u4e00\u4e0b\u8fd9\u5f20\u56fe\uff1a<\/p>\n<ol>\n<li>\u7528\u6237\u7684\u539f\u59cb query \u662f \u201cmichal jrdan\u201d<\/li>\n<li>Query Correction \u6a21\u5757\u8fdb\u884c\u62fc\u5199\u7ea0\u9519\u540e\u7684\u7ed3\u679c\u4e3a\uff1a\u201cMichael Jordan\u201d<\/li>\n<li>Query Suggestion \u6a21\u5757\u8fdb\u884c\u4e0b\u62c9\u63d0\u793a\u7684\u7ed3\u679c\u4e3a\uff1a\u201cMichael Jordan berkley\u201d\u548c \u201cMichael Jordan NBA\u201d\uff0c\u5047\u8bbe\u7528\u6237\u9009\u62e9\u4e86\u201cMichael Jordan berkley\u201d<\/li>\n<li>Query Expansion \u6a21\u578b\u8fdb\u884c\u67e5\u8be2\u6269\u5c55\u540e\u7684\u7ed3\u679c\u4e3a\uff1a\u201cMichael Jordan berkley\u201d\u548c \u201cMichael I. Jordan berkley\u201d<\/li>\n<li>Query Classification \u6a21\u5757\u8fdb\u884c\u67e5\u8be2\u5206\u7c7b\u7684\u7ed3\u679c\u4e3a\uff1aacademic<\/li>\n<li>\u6700\u540e\u8bed\u4e49\u6807\u7b7e\uff08Semantic Tagging\uff09\u6a21\u5757\u8fdb\u884c\u547d\u540d\u5b9e\u4f53\u8bc6\u522b\u3001\u5c5e\u6027\u8bc6\u522b\u540e\u7684\u7ed3\u679c\u4e3a\uff1a[Michael Jordan: \u4eba\u540d][berkley:location]:academic<\/li>\n<\/ol>\n<p>\u4e0b\u9762\u6211\u4eec\u9010\u4e00\u7684\u6765\u770b\u770b\u8fd9\u9762\u8fd9\u4e9b\u6a21\u5757\u5185\u90e8\u7ec6\u8282\u3002<\/p>\n<p>\u9996\u5148\u770b\u4e00\u4e0b Query Correction \u6a21\u5757\uff0c\u4e5f\u5373\u67e5\u8be2\u7ea0\u9519\u6a21\u5757\u3002<\/p>\n<p>\u5bf9\u4e8e\u82f1\u6587\uff0c\u6700\u57fa\u672c\u7684\u8bed\u4e49\u5143\u7d20\u662f\u5355\u8bcd\uff0c\u56e0\u6b64\u62fc\u5199\u9519\u8bef\u4e3b\u8981\u5206\u4e3a\u4e24\u79cd\uff1a\u4e00\u79cd\u662f Non-word Error\uff0c\u6307\u5355\u8bcd\u672c\u8eab\u5c31\u662f\u62fc\u9519\u7684\uff0c\u6bd4\u5982\u5c06\u201chappy\u201d\u62fc\u6210\u201chbppy\u201d\uff0c\u201chbppy\u201d\u672c\u8eab\u4e0d\u662f\u4e00\u4e2a\u8bcd\u3002\u53e6\u5916\u4e00\u79cd\u662f Real-word Error\uff0c\u6307\u5355\u8bcd\u867d\u62fc\u5199\u6b63\u786e\u4f46\u662f\u7ed3\u5408\u4e0a\u4e0b\u6587\u8bed\u5883\u786e\u662f\u9519\u8bef\u7684\uff0c\u6bd4\u5982\u201ctwo eyes\u201d\u5199\u6210\u201ctoo eyes\u201d\uff0c\u201ctoo\u201d\u5728\u8fd9\u91cc\u662f\u660e\u663e\u9519\u8bef\u7684\u62fc\u5199\u3002<\/p>\n<p>\u800c\u5bf9\u4e8e\u4e2d\u6587\uff0c\u6700\u5c0f\u7684\u8bed\u4e49\u5355\u5143\u662f\u5b57\uff0c\u5f80\u5f80\u4e0d\u4f1a\u51fa\u73b0\u9519\u5b57\u7684\u60c5\u51b5\uff0c\u56e0\u4e3a\u73b0\u5728\u6bcf\u4e2a\u6c49\u5b57\u51e0\u4e4e\u90fd\u662f\u901a\u8fc7\u8f93\u5165\u6cd5\u8f93\u5165\u8bbe\u5907\uff0c\u4e0d\u50cf\u624b\u5199\u6c49\u5b57\u4e5f\u8bb8\u4f1a\u51fa\u9519\u3002\u867d\u7136\u6c49\u5b57\u53ef\u4ee5\u5355\u5b57\u6210\u8bcd\uff0c\u4f46\u662f\u4e24\u4e2a\u6216\u4ee5\u4e0a\u7684\u6c49\u5b57\u7ec4\u5408\u6210\u7684\u8bcd\u5374\u662f\u66f4\u5e38\u89c1\u7684\u8bed\u4e49\u5143\u7d20\uff0c\u8fd9\u79cd\u7ec4\u5408\u5e26\u6765\u4e86\u7c7b\u4f3c\u82f1\u6587\u7684 Non-word Error\uff0c\u6bd4\u5982\u201c\u5927\u6570\u636e\u201d\u5199\u6210\u201c\u5927\u6811\u636e\u201d\uff0c\u867d\u7136\u6bcf\u4e2a\u5b57\u662f\u5bf9\u7684\uff0c\u4f46\u662f\u6574\u4f53\u5374\u4e0d\u662f\u4e00\u4e2a\u8bcd\uff0c\u4e5f\u5c31\u662f\u6240\u8c13\u7684\u522b\u5b57\u3002<\/p>\n<p><strong>query \u7ea0\u9519\u7684\u5177\u4f53\u65b9\u6848\u6709\uff1a<\/strong><\/p>\n<ol>\n<li>\u57fa\u4e8e\u7f16\u8f91\u8ddd\u79bb<\/li>\n<li>\u57fa\u4e8e\u566a\u58f0\u4fe1\u9053\u6a21\u578b<\/li>\n<\/ol>\n<p><strong>1. \u7f16\u8f91\u8ddd\u79bb\u7684\u65b9\u6cd5<\/strong><\/p>\n<p>\u7f16\u8f91\u8ddd\u79bb\u5305\u62ec\u5220\u9664\uff08del\uff09\u3001\u589e\u52a0\uff08ins\uff09\u3001\u66ff\u6362\uff08sub\uff09\u548c\u98a0\u5012\uff08trans\uff09\u56db\u79cd\u8f6c\u79fb\u64cd\u4f5c\uff0c\u5bf9\u5e94\u56db\u79cd\u8f6c\u79fb\u77e9\u9635\u3002\u8fd9\u4e9b\u8f6c\u79fb\u77e9\u9635\u7684\u6982\u7387\u53ef\u4ee5\u901a\u8fc7\u5bf9\u8bed\u6599\u5e93\u7edf\u8ba1\u5927\u91cf\u7684\u6b63\u786e\u8bcd\u548c\u9519\u8bef\u8bcd\u5bf9\u6765\u5f97\u5230\u3002<\/p>\n<p>\u4e2d\u6587\u8bcd\u8bed\u7684\u7f16\u8f91\u8ddd\u79bb\u8f6c\u6362\u5b58\u5728\u8f83\u5927\u7684\u8f6c\u4e49\u98ce\u9669\u3002\u6bd4\u5982\uff0c\u96f7\u950b -&gt;\u96f7\u5cf0\u5854\u3002\u4e2d\u6587\u7ea0\u9519\u901a\u5e38\u4ee5\u62fc\u97f3\u4e3a\u57fa\u7840\uff0c\u7f16\u8f91\u8ddd\u79bb\u7b49\u5176\u4ed6\u65b9\u5f0f\u4e3a\u8f85\u7684\u7b56\u7565\u3002<\/p>\n<p><strong>2. \u57fa\u4e8e\u566a\u58f0\u4fe1\u9053\u6a21\u578b\u7684\u7ea0\u9519\u65b9\u6cd5<\/strong><\/p>\n<p>\u566a\u58f0\u4fe1\u9053\u6a21\u578b\uff08Noisy Channel Model\uff09\u6700\u65e9\u662f\u9999\u519c\u4e3a\u4e86\u6a21\u578b\u5316\u4fe1\u9053\u7684\u901a\u4fe1\u95ee\u9898\uff0c\u5728\u4fe1\u606f\u71b5\u6982\u5ff5\u4e0a\u63d0\u51fa\u7684\u6a21\u578b\uff0c\u76ee\u6807\u662f\u4f18\u5316\u566a\u58f0\u4fe1\u9053\u4e2d\u4fe1\u53f7\u4f20\u8f93\u7684\u541e\u5410\u91cf\u548c\u51c6\u786e\u7387\u3002\u5bf9\u4e8e\u81ea\u7136\u8bed\u8a00\u5904\u7406\u800c\u8a00\uff0c\u566a\u58f0\u4fe1\u9053\u6a21\u578b\u5982\u4e0b\u56fe\uff0c\u5176\u4e2d I \u8868\u793a\u8f93\u5165\uff0cO \u8868\u793a\u7ecf\u8fc7\u566a\u58f0\u4fe1\u9053\u540e\u7684\u8f93\u51fa\uff0cI\u2019\u8868\u793a\u7ecf\u8fc7\u89e3\u7801\u540e\u6700\u6709\u53ef\u80fd\u7684\u8f93\u5165\u3002<\/p>\n<p>\u81ea\u7136\u8bed\u8a00\u5904\u7406\u4e2d\u7684\u673a\u5668\u7ffb\u8bd1\uff0c\u8bcd\u6027\u6807\u6ce8\uff0c\u8bed\u97f3\u8bc6\u522b\u7b49\u591a\u4e2a\u95ee\u9898\u90fd\u53ef\u4ee5\u4f7f\u7528\u4fe1\u9053\u566a\u58f0\u6a21\u578b\u6765\u89e3\u51b3\uff0c\u5bf9\u4e8e\u7ea0\u9519\u95ee\u9898\u4e5f\u53ef\u4ee5\u4f7f\u7528\u4fe1\u9053\u566a\u58f0\u6a21\u578b\u6765\u89e3\u51b3\uff0c\u76f8\u5e94\u7684\u6c42\u89e3\u95ee\u9898\u53ef\u4ee5\u7528\u516c\u5f0f\u8868\u8fbe\uff1a<\/p>\n<p>\u5176\u4e2d p(x|w) \u662f\u6b63\u786e\u7684\u8bcd\u7f16\u8f91\u6210\u4e3a\u9519\u8bef\u8bcd x \u7684\u8f6c\u79fb\u6982\u7387\uff0cp(w) \u662f\u6b63\u786e\u8bcd\u7684\u6982\u7387\uff0cp(x) \u662f\u9519\u8bef\u8bcd\u7684\u6982\u7387\u3002<\/p>\n<p>\u566a\u58f0\u4fe1\u9053\u6a21\u578b\u548c Non-word \u7ea0\u9519\u7c7b\u4f3c\uff0c\u53ea\u662f\u8ba1\u7b97\u76ee\u6807\u4ece\u67d0\u4e2a\u5019\u9009\u8bcd\u7684\u6700\u5927\u6982\u7387\u53d8\u6210\u4e0d\u540c\u4f4d\u7f6e\u5019\u9009\u8bcd\u7ec4\u5408\u5f62\u6210\u7684\u53e5\u5b50 p(s) \u7684\u6700\u5927\u6982\u7387\uff0c\u8fd9\u4e2a\u95ee\u9898\u53ef\u4ee5\u4f7f\u7528 HMM\uff08Hidden Markov model\uff0c\u9690\u9a6c\u5c14\u53ef\u592b\u6a21\u578b\uff09\u6c42\u89e3\u3002<\/p>\n<p>\u9664\u4e86\u4e0a\u8ff0 2 \u79cd\u65b9\u6cd5\uff0c\u771f\u5b9e\u7684\u573a\u666f\u4e2d\uff0c\u8fd8\u4f1a\u7ed3\u5408\u641c\u7d22\u65e5\u5fd7\u7684 session \u5206\u6790\u548c\u70b9\u51fb\u6a21\u578b\u6765\u8fdb\u884c\u7ea0\u9519\u7ed3\u679c\u7684\u6392\u5e8f\u8c03\u6574\u3002\u8fd9\u91cc\u4e0d\u518d\u5c55\u5f00\u3002<\/p>\n<p>\u73b0\u5728\u6211\u4eec\u518d\u770b\u770b Query Suggest \u6a21\u5757,Query Suggest\uff0c\u4e5f\u5373\u8f93\u5165\u4e0b\u62c9\u63d0\u793a\uff0c\u6839\u636e\u7528\u6237\u8f93\u5165\u7684\u67e5\u8be2\u8bcd\u524d\u7f00\u81ea\u52a8\u63d0\u793a\u7528\u6237\u6700\u6709\u53ef\u80fd\u8f93\u5165\u7684\u5b8c\u6574\u67e5\u8be2\u8bcd\u5217\u8868\u3002<\/p>\n<p><strong>\u8fd9\u91cc\u6d89\u53ca\u51e0\u4e2a\u95ee\u9898<\/strong>\uff1a<\/p>\n<ol>\n<li>Suggest \u8bcd\u6761\u6765\u4ece\u54ea\u91cc\u6765<\/li>\n<li>\u5982\u4f55\u6839\u636e\u5f53\u524d\u7684\u8f93\u5165\u5feb\u901f\u5339\u914d\u5230\u5019\u9009 suggest \u8bcd\u6761\u5217\u8868<\/li>\n<li>\u5982\u4f55\u5bf9\u5019\u9009 suggest \u8bcd\u6761\u5217\u8868\u8fdb\u884c\u6392\u5e8f<\/li>\n<\/ol>\n<p><img decoding=\"async\" src=\"https:\/\/www.fagao.me\/images\/uploads\/2018_001.jpg\" alt=\"\"><\/p>\n<p>suggest \u8bcd\u6761\u901a\u5e38\u4e3b\u8981\u6765\u81ea\u7528\u6237\u641c\u7d22\u5386\u53f2 query log\uff0c\u4f46\u5b58\u5728\u6570\u636e\u51b7\u542f\u52a8\u7684\u95ee\u9898\uff0c\u5f00\u59cb\u65f6\u7f3a\u5c11 query log \u65f6\u5982\u4f55\u5904\u7406\uff1f\u5bf9\u4e8e\u4e00\u4e9b\u5782\u76f4\u7684\u5e94\u7528\u573a\u666f\uff0c\u6bd4\u5982\u5c0f\u8bf4\u641c\u7d22\u4e2d\uff0csuggest \u8bcd\u6761\u4e5f\u53ef\u4ee5\u662f\u4f5c\u54c1\u7684\u6807\u9898\u3001\u6807\u7b7e\u3001\u4f5c\u5bb6\u540d\u7b49\uff0c\u7535\u5546\u641c\u7d22\u4e2d\u53ef\u4ee5\u662f\u54c1\u724c\u8bcd\u5e93\u3001\u4ea7\u54c1\u5217\u8868\u7b49\u3002<\/p>\n<p>\u5bf9\u4e8e suggest \u8bcd\u6761\u5217\u8868\u7684\u5b58\u50a8\u7ed3\u6784\u4e0e\u5feb\u901f\u5339\u914d\uff0c\u5982\u679c suggest \u8bcd\u6761\u5217\u8868\u4e0d\u662f\u5f88\u5927\uff0cTrie \u6811\uff08\u53c8\u79f0\u5b57\u5178\u6811\uff09\u662f\u4e00\u4e2a\u4e0d\u9519\u7684\u9009\u62e9\uff0c\u7528 Trie \u6811\u5b9e\u73b0\u7684\u4e3b\u8981\u4f18\u70b9\u662f\u5229\u7528\u5b57\u7b26\u4e32\u7684\u516c\u5171\u524d\u7f00\u6765\u964d\u4f4e\u67e5\u8be2\u65f6\u95f4\u7684\u5f00\u9500\u4ee5\u8fbe\u5230\u63d0\u9ad8\u6548\u7387\u7684\u76ee\u7684\uff0c\u5b9e\u73b0\u4e5f\u6bd4\u8f83\u7b80\u5355\uff0c\u4f46\u7f3a\u70b9\u662f\u8282\u70b9\u6570\u589e\u52a0\u5230\u4e00\u5b9a\u7a0b\u5ea6\u5185\u5b58\u5f00\u9500\u4f1a\u6210\u4e3a\u74f6\u9888\u3002\u5982\u679c suggest \u8bcd\u6761\u5217\u8868\u5f88\u5927\uff0c\u53ef\u4ee5\u9009\u62e9 Ternary Tree(\u53c8\u79f0\u4e09\u53c9\u641c\u7d22\u6811)\uff0c \u4e09\u53c9\u641c\u7d22\u6811\u5bf9 Trie \u7684\u5185\u5b58\u95ee\u9898\uff08\u7a7a\u7684\u6307\u9488\u6570\u7ec4\uff09\u8fdb\u884c\u4e86\u4e13\u95e8\u4f18\u5316\uff0c\u5177\u4f53\u7ec6\u8282\u5927\u5bb6\u53ef\u4ee5 google\uff0c\u8fd9\u91cc\u4e0d\u518d\u6df1\u5165\u3002<\/p>\n<p>Suggest \u5019\u9009\u8bcd\u7684\u6392\u5e8f\u901a\u5e38\u6839\u636e\u5019\u9009\u8bcd\u9879\u7684\u6574\u4f53\u70ed\u95e8\u6307\u6570\uff0c\u5373\u641c\u7d22\u7684\u591a\u7684\u6392\u5728\u524d\u9762\u3002\u5f53\u7136\u5b9e\u9645\u5e94\u7528\u573a\u666f\u4e2d\u7684\u6392\u5e8f\u4f1a\u8003\u8651\u66f4\u591a\u7684\u6392\u5e8f\u56e0\u7d20\uff0c\u6bd4\u5982\u4e2a\u6027\u5316\u7684\u56e0\u7d20\uff0c\u5f53\u4e0b\u70ed\u95e8\u6307\u6570\u7b49\u56e0\u7d20\u3002<\/p>\n<p><strong>Query Expansion \u67e5\u8be2\u6269\u5c55\u6a21\u5757<\/strong><\/p>\n<p>\u67e5\u8be2\u8bcd\u6269\u5c55\u6280\u672f\u901a\u8fc7\u5c06\u4e0e\u7528\u6237\u67e5\u8be2\u8bcd\u76f8\u8fd1\u3001\u76f8\u5173\u7684\u8bcd\u6269\u5c55\u5230\u7528\u6237\u67e5\u8be2\u8bcd\u4e2d\u7684\u65b9\u6cd5, \u66f4\u51c6\u786e\u5730\u63cf\u8ff0\u7528\u6237\u7684\u4fe1\u606f\u9700\u6c42, \u53bb\u9664\u7528\u6237\u67e5\u8be2\u8bcd\u7684\u591a\u4e49\u6027, \u4ece\u800c\u66f4\u7cbe\u786e\u5730\u67e5\u8be2\u7528\u6237\u6240\u9700\u4fe1\u606f\u3002\u5728\u4fe1\u606f\u68c0\u7d22\u6280\u672f\u4e2d, \u67e5\u8be2\u8bcd\u6269\u5c55\u662f\u4e00\u79cd\u80fd\u591f\u6709\u6548\u63d0\u9ad8\u67e5\u8be2\u6548\u7387\u7684\u6280\u672f\u3002\u901a\u8fc7\u7528\u6237\u641c\u7d22\u65e5\u5fd7\u548c\u70b9\u51fb\u65e5\u5fd7\u53ef\u4ee5\u6316\u6398\u51fa\u67e5\u8be2\u6269\u5c55\u8bcd\u3002<\/p>\n<p>\u6211\u4eec\u5728\u5b9e\u8df5\u4e2d\u91c7\u7528\u4e00\u79cd\u57fa\u4e8e\u641c\u7d22\u65e5\u5fd7\u4f1a\u8bdd\u5c40\u90e8\u4e0a\u4e0b\u6587\u548c\u5168\u5c40\u4e0a\u4e0b\u6587\u4e3a\u8bed\u6599\u5e93\u4f7f\u7528 word2vec \u6784\u5efa skip-gram \u8bcd\u5411\u91cf\u6a21\u578b\uff0c\u6839\u636e\u8bcd\u5411\u91cf\u6a21\u578b\u53ef\u4ee5\u53d6\u5f97\u4e0e\u67e5\u8be2\u8bcd\u6700\u76f8\u4f3c\u7684\u524d N \u4e2a\u8bcd\u6784\u6210\u521d\u6b65\u7684\u76f8\u5173\u5019\u9009\u8bcd\u8868\uff0c\u7136\u540e\u518d\u5229\u7528 K \u8fd1\u90bb\u7b97\u6cd5\u4ece\u76f8\u5173\u8bcd\u5019\u9009\u8bcd\u8868\u9009\u53d6\u51fa\u8bed\u4e49\u6700\u76f8\u5173\u7684\u5019\u9009\u8bcd\u4f5c\u4e3a\u67e5\u8be2\u8bcd\u7684\u6269\u5c55\u8bcd\u3002<\/p>\n<p>\u641c\u7d22\u65e5\u5fd7\u4f1a\u8bdd\u5c40\u90e8\u4e0a\u4e0b\u6587\u662f\u6307\u4e0e\u5f53\u524d query \u5728\u540c\u4e00\u4e2a\u4f1a\u8bdd\u4e0a\u4e0b\u6587\u4e2d\u7684\u5171\u73b0 query, \u4e5f\u662f\u7528\u6237\u5bf9 query \u7684\u67e5\u8be2\u91cd\u6784\uff0c\u6bd4\u5982\u521d\u59cb query \u4e3a\u201c\u53d8\u5f62\u91d1\u521a\u201d\uff0c\u7528\u6237\u5728\u67e5\u8be2\u4f1a\u8bdd\u4e2d\u53ef\u80fd\u5c06 query \u53d8\u6362\u4e3a \u201c\u53d8\u5f62\u91d1\u521a\u7535\u5f71\u201d\u8fdb\u884c\u641c\u7d22\uff0c\u5219\u201c\u53d8\u5f62\u91d1\u521a\u7535\u5f71\u201d\u4e3a\u539f\u59cb query \u7684\u5c40\u90e8\u4e0a\u4e0b\u6587\u3002<\/p>\n<p>query \u7684\u5168\u5c40\u4e0a\u4e0b\u6587\u6316\u6398\u601d\u8def\uff1a<\/p>\n<p>\u6839\u636e\u67e5\u8be2\u8bcd\u548c\u67e5\u8be2\u6240\u70b9\u51fb\u7684\u7ed3\u679c\u6784\u5efa\u4e8c\u90e8\u56fe\uff0c\u5229\u7528\u968f\u673a\u6e38\u8d70\u6a21\u578b\u8ba1\u7b97\u51fa\u6bcf\u4e2a\u67e5\u8be2\u8bcd\u7684\u6587\u6863\u5206\u5e03\u4f5c\u4e3a\u67e5\u8be2\u8bcd\u7684\u67e5\u8be2\u5411\u91cf\uff0c\u518d\u5229\u7528 KL \u8ddd\u79bb\u6765\u8ba1\u7b97\u4e24\u67e5\u8be2\u5411\u91cf\u4e4b\u95f4\u7684\u76f8\u4f3c\u6027\u3002<\/p>\n<p><strong>Query Classification \u67e5\u8be2\u610f\u56fe\u5206\u7c7b\u6a21\u5757<\/strong><\/p>\n<p>\u901a\u5e38\u6709\u57fa\u4e8e\u89c4\u5219\u6a21\u677f\u7684\u5206\u7c7b\u65b9\u6cd5\u548c\u57fa\u4e8e\u673a\u5668\u5b66\u4e60\u7684\u5206\u7c7b\u65b9\u6cd5\u3002<\/p>\n<p>\u4e00\u79cd\u662f\u57fa\u4e8e\u89c4\u5219\u6a21\u677f\u7684\u5206\u7c7b\u65b9\u6cd5\uff0c\u8fd9\u79cd\u65b9\u6cd5\u6bd4\u8f83\u9002\u7528\u4e8e\u67e5\u8be2\u975e\u5e38\u7b26\u5408\u89c4\u5219\u7684\u7c7b\u522b\uff0c\u901a\u8fc7\u89c4\u5219\u89e3\u6790\u7684\u65b9\u5f0f\u6765\u83b7\u53d6\u67e5\u8be2\u7684\u610f\u56fe\u3002\u6bd4\u5982\uff1a\u4eca\u5929\u5929\u6c14\u600e\u4e48\u6837\uff0c \u53ef\u4ee5\u8f6c\u5316\u4e3a [\u65e5\u671f][\u5b9e\u4f53: \u5929\u6c14][\u8be2\u95ee\u8bcd: \u600e\u4e48\u6837]\u4e0a\u6d77\u5230\u66fc\u8c37\u7684\u673a\u7968\u4ef7\u683c\uff0c \u53ef\u4ee5\u8f6c\u5316\u4e3a [\u5730\u70b9] \u5230 [\u5730\u70b9][\u673a\u7968 \/ \u8f66\u7968 \/ \u706b\u8f66\u7968] \u4ef7\u683c<\/p>\n<p>\u8fd9\u79cd\u65b9\u6cd5\u7684\u5bf9\u6bd4\u8f83\u660e\u786e\u7684\u89c4\u5219\u6027\u5f3a\u7684\u65b9\u5f0f\u6709\u7cbe\u786e\u7684\u8bc6\u522b\u5ea6\uff0c\u7f3a\u70b9\u662f\u8986\u76d6\u5ea6\u4f4e\uff0c\u7528\u6237\u67e5\u8be2\u7a0d\u4f5c\u53d8\u6362\u53ef\u80fd\u5c31\u4e0d match \u4e86\uff0c\u53e6\u5916\u89c4\u5219\u7684\u53d1\u73b0\u548c\u5236\u5b9a\u4e3b\u8981\u9760\u4eba\u5de5\u8fdb\u884c\u3002<\/p>\n<p>\u53e6\u4e00\u79cd\u662f\u57fa\u4e8e\u673a\u5668\u5b66\u4e60\u5206\u7c7b\u7684\u65b9\u6cd5\u3002<\/p>\n<p>\u5982\u679c\u6709\u786e\u5b9a\u7684\u67e5\u8be2\u7c7b\u522b\u4f53\u7cfb\uff0c\u57fa\u4e8e\u673a\u5668\u5b66\u4e60\u7684\u67e5\u8be2\u610f\u56fe\u5206\u7c7b\u662f\u4e00\u4e2a\u4e0d\u9519\u7684\u9009\u62e9\uff0c\u53ef\u4ee5\u9009\u62e9 SVM \u4f5c\u4e3a\u5206\u7c7b\u5668\uff0c\u5173\u952e\u5728\u5206\u7c7b\u7279\u5f81\u7684\u9009\u62e9, \u8fd8\u6709\u8bad\u7ec3\u6837\u672c\u7684\u51c6\u786e\u6807\u6ce8\u3002<\/p>\n<p>\u8fd9\u4e2a\u548c\u6211\u4eec\u4e4b\u524d\u53c2\u52a0\u8fc7\u7684 2014 ACM CIKM \u7ade\u8d5b\u7684\u95ee\u9898\u7c7b\u4f3c\uff0c\u90a3\u5e74 CIKM \u7ade\u8d5b\u7684\u9898\u76ee\u662f\u81ea\u52a8\u8bc6\u522b\u7528\u6237\u7684\u67e5\u8be2\u610f\u56fe\uff08Query Intent Detection\uff0cQID\uff09\uff1a\u7ed9\u5b9a\u4e00\u6279\u6807\u6ce8\u8fc7\u7c7b\u522b\u7684\u641c\u7d22\u65e5\u5fd7\u5305\u62ec\u67e5\u8be2\u65e5\u5fd7\u548c\u70b9\u51fb\u65e5\u5fd7\u4f5c\u4e3a\u8bad\u7ec3\u6837\u672c\uff0c\u5176\u4e2d\u4e5f\u6709\u90e8\u5206\u672a\u6807\u6ce8\u7684\uff0c\u7c7b\u522b\u4e3a unknown\u3002<\/p>\n<p>\u5728\u7279\u5f81\u7684\u9009\u62e9\u65b9\u9762\uff0c\u9664\u4e86\u57fa\u672c\u7684 Query \u7684\u957f\u5ea6\u3001Query \u7684\u9891\u6b21\u3001Title \u7684\u957f\u5ea6\u3001Title \u7684\u9891\u6b21\u3001BM-25\u3001Query \u7684\u9996\u5b57\u3001\u5c3e\u5b57\u7b49\uff0c\u6211\u4eec\u901a\u8fc7\u5bf9 log session \u4e0a\u4e0b\u6587\u7684\u5206\u6790\uff0c\u8fdb\u884c\u4e86 Query \u95f4\u7279\u5f81\u8bcd\u6c47\u7684\u6316\u6398\uff0c\u8fd0\u7528\u4e86 query \u5728\u76f8\u540c session \u4e2d\u7684\u5171\u73b0\u5173\u7cfb\uff0c\u6316\u6398 query \u4e4b\u95f4\u7684\u5b50\u4e32\u5305\u542b\u5173\u7cfb\uff0cquery \u548c\u70b9\u51fb\u7684 title \u4e4b\u95f4\u7684\u6587\u672c\u7279\u5f81\u5173\u7cfb\u7b49\u3002<\/p>\n<p>\u5728\u5206\u7c7b\u6a21\u578b\u7684\u9009\u62e9\u65b9\u9762\uff0c\u6211\u4eec\u9009\u62e9\u4e86 Ensemble \u6846\u67b6\u3002Ensemble \u7684\u57fa\u672c\u601d\u60f3\u662f\u5145\u5206\u8fd0\u7528\u4e0d\u540c\u5206\u7c7b\u7b97\u6cd5\u5404\u79cd\u7684\u4f18\u52bf\uff0c\u53d6\u957f\u8865\u77ed\uff0c\u7ec4\u5408\u5f62\u6210\u4e00\u4e2a\u5f3a\u5927\u7684\u5206\u7c7b\u6846\u67b6\u3002\u4e0d\u8fc7 Ensemble \u4e0d\u662f\u7b80\u5355\u7684\u628a\u591a\u4e2a\u5206\u7c7b\u5668\u5408\u5e76\u8d77\u6765\u7ed3\u679c\uff0c\u6216\u8005\u7b80\u5355\u5c06\u5206\u7c7b\u7ed3\u679c\u6309\u56fa\u5b9a\u53c2\u6570\u7ebf\u6027\u53e0\u52a0 (\u4f8b\u5982\u4e0d\u662f a1 * ALGO1 + a2 * ALGO2 + a3 * ALGO3)\uff0c\u800c\u662f\u901a\u8fc7\u8bad\u7ec3 Ensemble \u6a21\u578b\uff0c\u6765\u5b9e\u73b0\u6700\u4f18\u7684\u7ec4\u5408\u3002<\/p>\n<p>\u5728 Ensemble \u6846\u67b6\u4e0b\uff0c\u6211\u4eec\u5206\u7c7b\u5668\u5206\u4e3a\u4e24\u4e2a Level: L1 \u5c42\u548c L2 \u5c42\u3002L1 \u5c42\u662f\u57fa\u7840\u5206\u7c7b\u5668\uff0cL2 \u5c42\u57fa\u4e8e L1 \u5c42\uff0c\u5c06 L1 \u5c42\u7684\u5206\u7c7b\u7ed3\u679c\u5f62\u6210\u7279\u5f81\u5411\u91cf\uff0c\u518d\u7ec4\u5408\u4e00\u4e9b\u5176\u4ed6\u7684\u7279\u5f81\u540e\uff0c\u5f62\u6210 L2 \u5c42\u5206\u7c7b\u5668\uff08\u5982 SVM\uff09\u7684\u8f93\u5165\u3002\u8fd9\u91cc\u9700\u8981\u7279\u522b\u7559\u610f\u7684\u662f\u7528\u4e8e L2 \u5c42\u7684\u8bad\u7ec3\u7684\u6837\u672c\u5fc5\u987b\u6ca1\u6709\u5728\u8bad\u7ec3 L1 \u5c42\u65f6\u4f7f\u7528\u8fc7\u3002<\/p>\n<p><strong>Semantic Tagging \u6a21\u5757<\/strong><\/p>\n<p>\u8fd9\u4e2a\u6a21\u5757\u4e3b\u8981\u662f\u5bf9 query \u4e2d\u7684\u547d\u540d\u5b9e\u4f53\u8fdb\u884c\u8bc6\u522b\uff0c\u6bd4\u5982\u5bf9\u7535\u5546\u641c\u7d22 query \u4e2d\u7684\u54c1\u724c\u8bcd\u3001\u4ea7\u54c1\u8bcd\u3001\u5c5e\u6027\u8bcd\u3001\u5730\u5740\u8fdb\u884c\u8bc6\u522b\u3002\u5bf9 query\uff0c\u7528\u4e00\u4e2a\u76f8\u5bf9\u51c6\u786e\u7684\u8bcd\u5178 (\u54c1\u724c\u8bcd \/ \u4ea7\u54c1\u8bcd \/ \u5c5e\u6027\u8bcd \/ \u5730\u5740\u8bcd) \u53bb\u6807\u6ce8\u8bed\u6599\u3002<\/p>\n<p>\u6bd4\u5982\u5bf9\u4e8e \u201d\u65b0\u897f\u5170\u5b89\u4f73\u5976\u7c89\u4e8c\u6bb5\u201c \u6807\u6ce8\u8bed\u6599\u5982\u4e0b\u6240\u793a\uff1a\u65b0 B-loc \u897f I-loc \u5170 I-loc \u5b89 B-brand \u4f73 I-brand \u5976 B-product \u7c89 I-product \u4e8c B-attr \u6bb5 I-attr\u5b9e\u4f53\u8bcd\u8bc6\u522b\u6a21\u578b\u53ef\u4ee5\u901a\u8fc7 crf \u6765\u8fdb\u884c\u8bad\u7ec3\u3002<\/p>\n<p>\u81f3\u6b64\uff0c\u7b2c\u4e8c\u90e8\u5206 \u5982\u4f55\u8bc6\u522b\u7528\u6237\u641c\u7d22\u610f\u56fe \u4e5f\u8bb2\u5b8c\u4e86\u603b\u7ed3\u4e00\u4e0b\uff0c\u6211\u4eec\u9996\u5148\u7b80\u5355\u8bf4\u660e\u4e86\u7528\u6237\u641c\u7d22\u610f\u56fe\u7684\u4e3b\u8981\u5206\u7c7b\uff1a\u5bfc\u822a\u7c7b\u3001\u4fe1\u606f\u7c7b\u3001\u8d44\u6e90\u7c7b\uff0c\u7136\u540e\u5bf9\u641c\u7d22\u610f\u56fe\u8bc6\u522b\u7684\u4e3b\u8981\u529f\u80fd\u6a21\u5757\u67e5\u8be2\u7ea0\u9519\u3001\u67e5\u8be2\u8bcd\u81ea\u52a8\u63d0\u793a\u3001\u67e5\u8be2\u6269\u5c55\uff0c\u67e5\u8be2\u81ea\u52a8\u5206\u7c7b\u3001\u8bed\u4e49\u6807\u7b7e\u7b49\u5b9e\u73b0\u601d\u8def\u5206\u522b\u8fdb\u884c\u4e86\u4ecb\u7ecd\u3002<\/p>\n<p><span>\u4e09\u3001\u8fbe\u89c2\u641c\u7d22\u610f\u56fe\u8bc6\u522b\u5f15\u64ce<\/span><\/p>\n<p>\u8fbe\u89c2\u901a\u8fc7 RESTAPI \u63a5\u53e3\u7684\u65b9\u5f0f\u5411\u5ba2\u6237\u63d0\u4f9b\u57fa\u4e8e\u516c\u6709\u4e91\u548c\u79c1\u6709\u4e91\u7684\u641c\u7d22\u670d\u52a1\u3002\u5176\u4e2d\u8bed\u4e49\u5206\u6790\u6a21\u5757\u5305\u542b\u4e86\u5bf9\u7528\u6237 query \u610f\u56fe\u7684\u79bb\u7ebf\u6316\u6398\u548c\u5728\u7ebf\u9884\u6d4b\u3002<\/p>\n<p>\u8fbe\u89c2 query \u610f\u56fe\u8bc6\u522b\u5f15\u64ce\u5185\u90e8\u7ec4\u5408\u4e86\u591a\u7c92\u5ea6\u5206\u8bcd\u3001\u591a\u6b67\u4e49\u5206\u8bcd\u3001\u547d\u540d\u5b9e\u4f53\u8bc6\u522b\u3001\u8bcd\u6c47\u95f4\u7d27\u5bc6\u5ea6\u5206\u6790\u3001\u8bcd\u6c47\u91cd\u8981\u6027\u5206\u6790\u3001\u540c\u4e49\u8fd1\u4e49\u8bcd\u5206\u6790\u3001\u62fc\u5199\u7ea0\u9519\u3001\u62fc\u97f3\u8f6c\u5199\u3001\u9ad8\u9891\u8bcd\u52a8\u6001\u7701\u7565\u5206\u6790\u3001\u9700\u6c42\u8bcd\u8bc6\u522b\u3001\u76f8\u5173\u641c\u7d22\u8bcd\u3001\u76f4\u8fbe\u641c\u7d22\u8bcd\u5206\u6790\u7b49\u591a\u4e2a\u529f\u80fd\u6a21\u5757\u3002<\/p>\n<p>\u5f15\u64ce\u5e95\u5c42\u57fa\u4e8e\u8fbe\u89c2\u7edf\u4e00\u7684\u6587\u672c\u8bed\u4e49\u6316\u6398\u7b97\u6cd5\u5e73\u53f0\uff1a<\/p>\n<p>\u8fbe\u89c2\u6587\u672c\u8bed\u4e49\u6316\u6398\u7b97\u6cd5\u5e73\u53f0\u662f\u4e00\u4e2a\u878d\u5408\u4e86\u591a\u79cd\u7b97\u6cd5\u7684\u96c6\u6210\u5b66\u4e60\u5e73\u53f0\uff0c\u65e2\u5305\u62ec\u7ecf\u5178\u7684 SVM\u3001LR\u3001RF\u3001LDA \u7b49\u7b97\u6cd5\uff0c\u4e5f\u5305\u62ec CNN\u3001RNN\u3001LSTM\u3001BILSTM \u7b49\u6df1\u5ea6\u5b66\u4e60\u7b97\u6cd5\u3002\u6bd4\u5982\u5728\u5b9e\u8df5\u4e2d\uff0c\u6211\u4eec\u5c1d\u8bd5\u5c06\u7ebf\u6027\u7edf\u8ba1\u6a21\u578b CRF \u548c\u795e\u7ecf\u7f51\u7edc\u7ed3\u6784 LSTM \u76f8\u878d\u5408\u7684\u65b9\u6cd5\uff0c\u5728 LSTM \u7684\u8f93\u51fa\u7aef\u5c06 softmax \u4e0e CRF \u7ed3\u5408\u8d77\u6765\uff0c\u4f7f\u7528 LSTM \u89e3\u51b3\u63d0\u53d6\u5e8f\u5217\u7684\u7279\u5f81\u95ee\u9898\uff0c\u4f7f\u7528 CRF \u6709\u6548\u5229\u7528\u4e86\u53e5\u5b50\u7ea7\u522b\u7684\u6807\u8bb0\u4fe1\u606f\uff0c\u53d6\u5f97\u4e86\u4e0d\u9519\u7684\u6548\u679c\u3002<\/p>\n<div class=\"articleend\"><\/div>\n","protected":false},"excerpt":{"rendered":"<p>\u4eba\u7c7b\u81ea\u8bde\u751f\u4ee5\u6765\u5c31\u4f34\u968f\u7740\u5404\u79cd\u4fe1\u606f\u7684\u751f\u4ea7\u548c\u83b7\u53d6\uff0c\u5982\u4eca\u8fd9\u4e2a\u4fe1\u606f\u7206\u70b8\u7684 DT \u65f6\u4ee3\uff0c\u4eba\u4eec\u66f4\u662f\u88ab\u5404\u79cd\u4fe1\u606f\u6240\u5305\u56f4\u3002\u6211\u4eec\u77e5\u9053 [&hellip;]<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[8],"tags":[183],"class_list":["post-114974","post","type-post","status-publish","format-standard","hentry","category-news","tag-183"],"_links":{"self":[{"href":"https:\/\/www.fagao.me\/p\/wp-json\/wp\/v2\/posts\/114974","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.fagao.me\/p\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.fagao.me\/p\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.fagao.me\/p\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.fagao.me\/p\/wp-json\/wp\/v2\/comments?post=114974"}],"version-history":[{"count":0,"href":"https:\/\/www.fagao.me\/p\/wp-json\/wp\/v2\/posts\/114974\/revisions"}],"wp:attachment":[{"href":"https:\/\/www.fagao.me\/p\/wp-json\/wp\/v2\/media?parent=114974"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.fagao.me\/p\/wp-json\/wp\/v2\/categories?post=114974"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.fagao.me\/p\/wp-json\/wp\/v2\/tags?post=114974"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}