{"id":551,"date":"2017-04-13T21:48:28","date_gmt":"2017-04-13T13:48:28","guid":{"rendered":"http:\/\/www.mrtblog.cn\/?p=551"},"modified":"2023-03-04T21:16:49","modified_gmt":"2023-03-04T13:16:49","slug":"%e5%bc%80%e5%9d%91%e4%bd%bf%e7%94%a8sklearn%e8%bf%9b%e8%a1%8c%e6%9c%ba%e5%99%a8%e5%ad%a6%e4%b9%a0%e4%bb%a5%e6%90%9c%e7%8b%97%e7%94%a8%e6%88%b7%e7%94%bb%e5%83%8f%e4%b8%ba%e4%be%8b","status":"publish","type":"post","link":"http:\/\/www.mrtblog.cn\/?p=551","title":{"rendered":"\u4f7f\u7528sklearn\u8fdb\u884c\u6587\u672c\u5206\u7c7b(\u4ee5\u641c\u72d7\u7528\u6237\u753b\u50cf\u4e3a\u4f8b)"},"content":{"rendered":"
<\/span> 1,607<\/span> Views<\/span><\/div>

\u524d\u8a00<\/h3>\n

\u641c\u72d7\u7528\u6237\u753b\u50cf\u6316\u6398\u662f2016CCF\u5927\u6570\u636e\u7ade\u8d5b\u7684\u9898\u76ee\uff0c\u90a3\u65f6\u5019\u5bf9\u6570\u636e\u6316\u6398\u4e86\u89e3\u4e0d\u591a\uff0c\u518d\u52a0\u4e0a\u9876\u7740\u4e24\u500d\u4e8e\u73b0\u5728\u7684\u8bfe\u7a0b\u91cf\u548c\u8bfe\u7a0b\u4f5c\u4e1a= =\uff0c\u82b1\u8d39\u5927\u7ea6\u4e00\u534a\u7684\u4f11\u606f\u65f6\u95f4\u7528JAVA\u5b9e\u73b0\u7528\u6237\u5206\u7c7b\uff0c\u4f46\u662f\u6548\u679c\u4e0d\u597d\uff0c\u53ea\u80fd\u8fbe\u523056%\u5de6\u53f3\uff08\u80fd\u8fdb\u590d\u8d5b\u7684\u5927\u7ea666%\uff09\u3002\u8003\u8651\u5230\u4e0d\u80fd\u628a\u6ca1\u7cbe\u529b\u4f5c\u4e3a\u501f\u53e3\uff0c\u5f53\u65f6\u5c31\u51b3\u5b9a\u6012\u8865\u57fa\u7840\uff0c\u82e6\u5b66py(\u4ea4\u6613)\u548cskl\uff0c\u501f\u52a9skl\u5bf9\u7b97\u6cd5\u91cd\u5199\u548c\u4f7f\u7528python\u5bf9\u4ee3\u7801\u91cd\u6784\uff0c\u53ea\u613f\u4e0b\u6b21\u53c2\u8d5b\u4e0d\u518d\u662f\u4e00\u53ea\u83dc\u9e1f\u3002<\/p>\n

\u56de\u987e\u6587\u672c\u5206\u7c7b\u7684\u57fa\u672c\u6b65\u9aa4\uff1a<\/p>\n

\u6570\u636e\u6e90\u83b7\u53d6->\u6570\u636e\u9884\u5904\u7406->\u7279\u5f81\u9009\u62e9->\u5206\u7c7b\u5668\u5206\u7c7b->\u4f18\u5316\u8c03\u6574->\u8f93\u51fa\u7ed3\u679c<\/p>\n

\u672c\u6587\u5c06\u6309\u7167\u6587\u672c\u5206\u7c7b\u7684\u57fa\u672c\u65b9\u6cd5\uff0c\u4e00\u6b65\u4e00\u6b65\u5b8c\u6210\u8bad\u7ec3\u8fc7\u7a0b\u3002<\/p>\n

\n

1.\u51c6\u5907\u5de5\u4f5c+\u6570\u636e\u6e90<\/h3>\n

\u7cfb\u7edf\uff1a4\u68388GB\u7684Windows8.1\u64cd\u4f5c\u7cfb\u7edf<\/p>\n

\u5de5\u5177\uff1apython2.7.3+sklearn+numpy+matplotlib+scipy+jieba<\/p>\n

\u6587\u672c\u7f16\u8f91\u5668\uff1aSublime Text 3+UltraEdit<\/p>\n

\u641c\u72d7\u7528\u6237\u753b\u50cf\u6316\u6398\u590d\u8d5b\u6570\u636e \uff1ahttp:\/\/pan.baidu.com\/s\/1jHAmP2u<\/a><\/p>\n

\u6570\u636e\u683c\u5f0f\u4e3a\uff1a\u7528\u6237id \u5e74\u9f84 \u6027\u522b \u5b66\u5386 \u641c\u7d22\u8bb0\u5f55\u82e5\u5e72\u6761\uff08\u6570\u636e\u5747\u4ee5’\\t’\u5212\u5206\uff09<\/p>\n

\u76ee\u6807\uff1a\u6839\u636e\u5df2\u77e5\u7528\u6237\u7684\u4e09\u4e2a\u5c5e\u6027\u548c\u641c\u7d22\u8bb0\u5f55\uff0c\u63a8\u6d4b\u53ea\u542b\u6709\u641c\u7d22\u8bb0\u5f55\u7684\u7528\u6237\u7684\u4e09\u5c5e\u6027<\/p>\n

\n

2.\u6570\u636e\u9884\u5904\u7406<\/h3>\n

\u56e0\u4e3a\u6bd4\u8d5b\u5df2\u7ecf\u7ed3\u675f\uff0c\u7f3a\u5c11\u5bf9\u6d4b\u8bd5\u96c6\u7684\u8bc4\u5224\u65b9\u6cd5\uff0c\u6240\u4ee5\u672c\u6b21\u5b9e\u9a8c\u5c06\u8bad\u7ec3\u96c6(10W\u6570\u636e)\u63098:2\u62c6\u5206\uff08\u8003\u8651\u73b0\u5b9e\u60c5\u51b5\uff0c\u6570\u636e\u96c6\u6709\u90e8\u5206\u5c5e\u6027\u7f3a\u5931\u7684\u6837\u672c\uff0c\u5bf9\u4e8e\u8fd9\u90e8\u5206\u6837\u672c\u91c7\u7528\u76f4\u63a5\u820d\u5f03\u7684\u65b9\u6cd5\uff0c\u8be6\u7ec6\u89c1\u7b2c\u4e8c\u90e8\u5206\u6570\u636e\u6e05\u6d17\uff09\uff0c\u537380%\u7684\u8bad\u7ec3\u96c6\u4e3a\u65b0\u7684\u8bad\u7ec3\u96c6\uff0c20%\u7684\u8bad\u7ec3\u96c6\u4e3a\u65b0\u7684\u6d4b\u8bd5\u96c6\u3002<\/p>\n

(1)\u6570\u636e\u6e05\u6d17<\/h4>\n

user_tag_query.10W.TRAIN\u7f16\u7801\u65b9\u5f0f\u4e3aGB18030\uff0c\u5148\u5c06\u5176\u8f6c\u6362\u4e3aUTF-8\u7f16\u7801\uff0c\u53ef\u4f7f\u7528\u6587\u672c\u7f16\u8f91\u5668\u8f6c\u7801\uff0c\u8fd9\u91cc\u4f7f\u7528python\u5bf9\u6587\u4ef6\u8f6c\u7801\u3002<\/p>\n

\u521b\u5efaPrePareWork.py\u5199\u5165\u5982\u4e0b\u4ee3\u7801\uff1a<\/p>\n

\n#coding=utf-8\noriginFile = open("user_tag_query.10W.TRAIN","r")#\u521d\u59cb\u7684\u8bad\u7ec3\u96c6\u6587\u4ef6\ntrainFile = open("train.csv","w+")#\u6e05\u6d17\u540e\u7684\u8bad\u7ec3\u96c6\u6587\u4ef6\nline = originFile.readline()\ni = 0#\u8bb0\u5f55\u884c\u6570\ncount = 0#\u8bb0\u5f55\u6709\u6548\u6570\u636e\u6570\nwhile line:\n    i += 1\n    datas = line.split("\\t")\n    if datas[1] != '0' and datas[2] != '0' and datas[3] != '0':#\u53ea\u8bb0\u5f55\u6ca1\u6709\u6807\u7b7e\u7f3a\u5931\u7684\u7528\u6237\n        count += 1\n        context='\\t'.join(datas[0:4])+'\\t'+'\\t'.join(datas[4:]).decode('GB18030').encode('UTF-8')\n        trainFile.write(context)\n    line = originFile.readline()\nprint "records:",count\noriginFile.close()\n<\/code><\/pre>\n
 
\n\u8fd9\u6837\uff0c\u6211\u4eec\u5c31\u5f97\u5230\u5220\u9664\u4e86\u7f3a\u5931\u6807\u7b7e\u4e14\u7f16\u7801\u4e3aUTF-8\u7684\u8bad\u7ec3\u96c6train.csv:<\/p>\n
\"\"<\/a><\/p>\n
(2)\u5206\u8bcd\u53ca\u7b5b\u9009<\/h4>\n
\u4e2d\u6587\u8bed\u53e5\u4e0d\u540c\u4e8e\u82f1\u6587\uff0c\u5927\u90e8\u5206\u8bcd\u8bed\u7d27\u5bc6\u76f8\u63a5\uff08\u82f1\u6587\u5219\u5168\u90e8\u4f7f\u7528\u7a7a\u683c\u9694\u5f00\u5355\u8bcd\uff09\u3002\u4e3a\u4e86\u7ee7\u7eed\u4e0b\u4e00\u6b65\u7684\u7279\u5f81\u9009\u62e9\uff0c\u8fd9\u91cc\u6211\u4eec\u9700\u8981\u73b0\u5bf9\u8bed\u53e5\u5207\u5206\u6210\u8bcd\u8bed\uff0c\u5e76\u4ece\u4e2d\u5220\u9664\u5927\u91cf\u65e0\u610f\u4e49\u7684\u8bcd\u8bed\uff08\u5982\u2018\u7684\u2019\uff0c\u2018\u4e5f\u2019\uff0c\u2018\u662f\u2019\u7b49\u7b49\uff09\uff0c\u8fd9\u4e9b\u8bcd\u88ab\u79f0\u4e3a\u505c\u7528\u8bcd\uff0c\u5373\u5bf9\u7528\u6237\u7684\u5206\u7c7b\u6ca1\u6709\u592a\u5927\u610f\u4e49\u7684\u8bcd\u8bed\u3002<\/p>\n
\u4f46\u662f\u6839\u636eCCF\u7ade\u8d5b\u7b2c\u4e00\u540d<\/a>\u7684\u60f3\u6cd5\uff0c\u5b9e\u9645\u4e0a\u5728\u672c\u6b64\u5b9e\u9a8c\u4e2d\u7684\u7a7a\u683c\u3001\u201c\u4e4b\u201d\u7b49\u505c\u7528\u8bcd\u5bf9\u5206\u7c7b\u7ed3\u679c\u4f1a\u4ea7\u751f\u8f83\u5927\u5f71\u54cd\uff08\u5b66\u5386\u8f83\u9ad8\u7684\u4eba\u5bf9\u641c\u7d22\u5f15\u64ce\u7684\u4f7f\u7528\u66f4\u52a0\u5a34\u719f\uff0c\u6240\u4ee5\u5bf9\u7a7a\u683c\u548c\u2018\uff0c\u2019\u7684\u4f7f\u7528\u6b21\u6570\u9ad8\u4e8e\u5176\u4ed6\u4eba\uff1b\u5e74\u9f84\u8f83\u4f4e\u7684\u4eba\u5bf9\u90e8\u5206\u5c0f\u8bf4\u5982\u7384\u5e7b\u5c0f\u8bf4\u60c5\u6709\u72ec\u949f\uff0c\u641c\u7d22\u8bb0\u5f55\u4f1a\u5305\u542b\u5927\u91cf\u5982\u201c\u4e4b\u201d\uff08\u7f51\u6e38\u4e4bXXX\u5c0f\u8bf4\uff09\u7684\u8bcd\u8bed\uff1b\u7b49\u7b49\uff09\u3002\u4e3a\u4e86\u7b80\u5355\u8d77\u89c1\uff0c\u8fd9\u91cc\u4f7f\u7528jieba\u5206\u8bcd\u540e\uff0c\u53ea\u4fdd\u7559\u52a8\u8bcd\u548c\u540d\u8bcd\u4ee5\u53ca\u7edf\u8ba1\u7a7a\u683c\u6570\uff0c\u5220\u53bb\u5176\u4ed6\u6240\u6709\u8bcd\u8bed<\/p>\n
\u521b\u5efaDelWord.py\u6587\u4ef6\u5e76\u5199\u5165\u5982\u4e0b\u4ee3\u7801\uff1a<\/p>\n
\n#encoding=utf-8\nimport jieba\nimport jieba.posseg\nimport sys\nreload(sys)\nsys.setdefaultencoding('utf8')\noriginFile = open("train.csv","r")#\u521d\u59cb\u7684\u8bad\u7ec3\u96c6\u6587\u4ef6\ntrainFile = open("newTrain.csv","w")#\u5904\u7406\u540e\u7684\u8bad\u7ec3\u96c6\u6587\u4ef6\nline = originFile.readline()\nusefulWord = ['n','v']#\u4fdd\u5b58\u8bcd\u6027\nq = 0#\u884c\u6570\n\nwhile line:\n    if q % 1 == 0:\n    print 'deal with line:', q\n    q += 1\n    inputStr = line.split("\\t")\n    count = 0#\u8bb0\u5f55\u7a7a\u683c\n    userSaveWord = []#\u8bb0\u5f55\u7528\u6237\u7684\u6709\u6548\u5355\u8bcd\n    context=' '.join(inputStr[0:4])#\u4fdd\u5b58\u7528\u6237\u6807\u7b7e\u4fe1\u606f\n    for i in inputStr[4:]:\n        words = jieba.posseg.cut(i)\n        for w in words:\n            if w.word == ' ':\n                count += 1\n                continue\n            for u in usefulWord:\n                if w.flag == u:\n                    userSaveWord.append(w.word.encode('UTF-8'))\n                    break\n                else:\n                    continue\n\n    context += ' ' + str(count) + ' ' + ' '.join(userSaveWord[:]) + '\\n'\n    trainFile.write(context)\n    line = originFile.readline()\n<\/code><\/pre>\n
\u5206\u8bcd\u65f6\u95f4\u5f88\u957f\uff0c\u5728\u7f51\u4e0a\u67e5\u627e\u76f8\u5173\u8d44\u6599\u53d1\u73b0windows\u4e0b\u4e0d\u80fd\u4f7f\u7528jieba\u7684\u5e76\u884c\u5206\u8bcd\u6a21\u5757\uff0c\u6240\u4ee5\u8fd9\u91cc\u6682\u4e0d\u8003\u8651\u5982\u4f55\u66f4\u6709\u6548\u7684\u63d0\u9ad8\u5206\u8bcd\u901f\u5ea6\uff0c\u8010\u5fc3\u7b49\u5f85\u5206\u8bcd\u5b8c\u6210\uff08\u8fd9\u4e00\u6b21\u53ef\u80fd\u4e0d\u4f1a\u4f7f\u7528\u5168\u90e8\u5927\u7ea69w\u7684\u7528\u6237\uff0c\u4f46\u662f\u53ef\u4ee5\u4fdd\u5b58\u4e0b\u6765\u4e0b\u6b21\u4f7f\u7528\uff09
\n\"\"<\/a><\/p>\n
\n
3.\u7279\u5f81\u9009\u62e9<\/h3>\n
\u6309\u7167\u57fa\u672c\u7684\u6587\u672c\u5206\u7c7b\u65b9\u6cd5\uff0c\u6211\u4eec\u53ea\u9700\u8981\u7edf\u8ba1\u6240\u6709\u51fa\u73b0\u8fc7\u7684\u8bcd\u6c47\uff0c\u5e76\u8bbe\u8ba1\u4e00\u4e2a\u8bcd\u888b\uff08\u4fdd\u5b58\u6240\u6709\u4e0d\u91cd\u590d\u5355\u8bcd\uff09\uff0c\u6839\u636e\u8bcd\u888b\u6c42\u5f97\u6bcf\u4e2a\u7528\u6237\u7684\u4e00\u4e2a\u5173\u4e8e\u8bcd\u8bed\u51fa\u73b0\u6b21\u6570\u7684\u5411\u91cf\uff0c\u6240\u6709\u7528\u6237\u7ec4\u6210\u4e3a\u8bad\u7ec3\u96c6\uff0c\u653e\u5165\u5206\u7c7b\u5668\u5206\u7c7b\u9884\u6d4b\u7ed3\u679c\u3002<\/p>\n
\u867d\u7136\u5728\u4e4b\u524d\u6570\u636e\u9884\u5904\u7406\u6211\u4eec\u5df2\u7ecf\u5220\u53bb\u4e86\u5927\u90e8\u5206\u7684\u505c\u7528\u8bcd\uff0c\u4f46\u662f\u73b0\u5728\u662f\u5426\u5c31\u5df2\u7ecf\u9002\u5408\u5f00\u59cb\u5206\u7c7b\u5904\u7406\u4e86\u5462\uff1f<\/p>\n
\u5220\u9664\u505c\u7528\u8bcd\u540e\u7684\u8bcd\u888b\u5305\u542b\u8bcd\u8bed\u5927\u7ea625w\uff0c\u7528\u6237\u6570\u91cf\u5927\u7ea69w\uff0c\u5982\u679c\u5c06\u6240\u6709\u8bcd\u8bed\u4f5c\u4e3a\u5411\u91cf\u7ef4\u5ea6\u4e22\u5165\u5206\u7c7b\u5668\uff0c\u90a3\u4e48\u6211\u4eec\u5c06\u5f97\u5230\u4e00\u4e2a25w*9w\u7684\u5de8\u5927\u7a00\u758f\u77e9\u9635\uff0c\u4e0d\u4ec5\u5904\u7406\u4e0a\u4f1a\u5f88\u590d\u6742\uff0c\u8fd8\u4f1a\u4f7f\u7ed3\u679c\u7ecf\u5ea6\u964d\u4f4e<\/p>\n
\u73b0\u5728\u5148\u653e\u4e0b\u8fd9\u4e9b\u4e0d\u7ba1\uff0c\u6211\u4eec\u53ef\u4ee5\u5f00\u59cb\u505a\u4e9b\u6709\u8da3\u7684\u4e8b\uff08\u4e5f\u662f\u505a\u6570\u636e\u6316\u6398\u5fc5\u4e0d\u53ef\u5c11\u7684\u4e00\u4ef6\u4e8b\u2014\u2014\u2014\u2014\u6570\u636e\u5206\u6790\uff09\uff0c\u6211\u4eec\u662f\u5426\u53ef\u4ee5\u6839\u636e\u6837\u672c\u7684\u5206\u5e03\u60c5\u51b5\uff0c\u5236\u5b9a\u76f8\u5e94\u7684\u5bf9\u7b56\u5462<\/p>\n
\u521b\u5efaCountInfo.py\u5199\u5165\u5982\u4e0b\u4ee3\u7801\uff1a<\/p>\n
\n#coding=utf-8\nfrom numpy import *\noriginFile = open("newTrain.csv","r")#\u521d\u59cb\u7684\u8bad\u7ec3\u96c6\u6587\u4ef6\ntrainFile = open("info.csv","w")#\u6e05\u6d17\u540e\u7684\u8bad\u7ec3\u96c6\u6587\u4ef6\n\nage = [0,0,0,0,0,0]\nsex = [0,0]\nedu = [0,0,0,0,0,0]\n\nline = originFile.readline()\nwhile line:\n    datas = line.split(" ")\n    a = datas[1];\n    s = datas[2];\n    e = datas[3]\n    #print a,s,e\n    age[int(a) - 1] += 1;sex[int(s) - 1] += 1;edu[int(e) - 1] += 1\n    line = originFile.readline()\n\ntrainFile.write(str(age)+"\\n")\ntrainFile.write(str(sex)+"\\n")\ntrainFile.write(str(edu)+"\\n")\n\nprint "age:",age\nprint "sex:",sex\nprint "edu:",edu\n<\/code><\/pre>\n
\u8fd0\u884c\u540e\u5c06\u4f1a\u5f97\u5230\u7ed3\u679c\uff0c\u663e\u793a\u4e0d\u540c\u5c5e\u6027\u5728\u8bad\u7ec3\u96c6\u4e2d\u6bcf\u4e2a\u7c7b\u522b\u7684\u51fa\u73b0\u6b21\u6570\uff1a<\/p>\n
\nage: [35965, 24785, 16561, 8967, 2039, 175]\nsex: [51540, 36952]\nedu: [354, 556, 18708, 27956, 36852, 4066]\n[Finished in 4.4s]\n<\/code><\/pre>\n
\u901a\u8fc7\u7b80\u5355\u7684\u5206\u6790\uff0c\u6211\u4eec\u5c31\u53ef\u4ee5\u505a\u51fa\u4e00\u4e9b\u51b3\u5b9a\uff0c\u751a\u81f3\u5728\u4e0d\u7528\u540e\u9762\u7684\u5206\u7c7b\u65b9\u6cd5\u7684\u60c5\u51b5\u4e0b\uff0c\u76f4\u63a5\u9884\u6d4b\u7ed3\u679c
\n\u6837\u672c\u603b\u6570\u5927\u7ea69\u4e07\uff0c\u73b0\u5728\u6211\u4eec\u5c06\u6240\u6709\u9884\u6d4b\u5c5e\u6027\u8bbe\u4e3aage:1,sex:1,edu:5,\u9884\u6d4b\u6b63\u786e\u7387\u5206\u522b\u4e3a40%,57%,41%\u3002\u662f\u4e0d\u662f\u5f88\u60ca\u559c\uff0c\u5e73\u5747\u4e0b\u6765\u5df2\u7ecf\u670950%\u7684\u51c6\u786e\u7387\u4e86\u3002\u53e6\u5916\uff0c\u8fd9\u4e2a\u6bd4\u4f8b\u53ef\u4f5c\u5148\u9a8c\u6982\u7387\u7528\u4e8e\u5206\u6790\u9884\u6d4b\u7ed3\u679c<\/p>\n
\u63a5\u4e0b\u6765\uff0c\u6211\u4eec\u56de\u5230\u521a\u521a\u7684\u95ee\u9898\uff0c\u5982\u4f55\u89e3\u51b3\u7ef4\u6570\u707e\u96be\u3002\u4e3a\u4e86\u51cf\u5c11\u8bcd\u888b\u4e2d\u603b\u5355\u8bcd\u7684\u6b21\u6570\uff0c\u6211\u4eec\u5e0c\u671b\u53ef\u4ee5\u627e\u5230\u5177\u6709\u4ee3\u8868\u6027\u7684\u8bcd\u8bed\uff0c\u90a3\u4e48\u53ea\u5728\u6781\u5c11\u6570\u4eba\u4e2d\u51fa\u73b0\u7684\u8bcd\u8bed\u662f\u5426\u6709\u4ee3\u8868\u610f\u4e49\uff1f\u51fa\u73b0\u5728\u7edd\u5927\u591a\u6570\u4eba\u4e2d\u7684\u8bcd\u8bed\u662f\u5426\u6709\u4ee3\u8868\u610f\u4e49\uff1f
\n\u8fd9\u5c31\u5df2\u7ecf\u57fa\u672c\u6d89\u53ca\u5230\u4e86\u7279\u5f81\u964d\u7ef4\u7406\u5ff5\u4e86\uff0c\u964d\u7ef4\u5904\u7406\u7684\u65b9\u6cd5\u6709\u5f88\u591a\uff0c\u8fd9\u91cc\u4f7f\u7528\u5361\u65b9\u6821\u9a8c\u7684\u65b9\u5f0f\uff0c\u597d\u5728sklearn\u4e5f\u4e3a\u6211\u4eec\u51c6\u5907\u4e86chi\u65b9\u6cd5\u3002<\/p>\n
\u901a\u8fc7\u5361\u65b9\u964d\u7ef4\u540e\u7684\u7528\u6237\u77e9\u9635\u4f5c\u4e3a\u8f93\u5165\uff0c\u518d\u7ecf\u8fc7\u5206\u7c7b\u5668svm\uff08\u9ed8\u8ba4\u53c2\u6570\uff09\u7684\u9884\u6d4b\u5c31\u53ef\u4ee5\u5f97\u5230\u521d\u6b65\u7ed3\u679c\u4e86\u3002\u8fd9\u91cc\u6211\u4eec\u5148\u5bf9\u8bad\u7ec3\u96c6\u524d1000\u7684\u6570\u636e\u63098:2\u5212\u5206\u5e76\u6d4b\u8bd5\u7ed3\u679c\uff0c\u540c\u65f6\u5bf9\u5361\u65b9\u6821\u9a8c\u7b5b\u9009\u540e\u7684\u7ef4\u5ea6\u904d\u5386\uff0c\u627e\u5230\u7ed3\u679c\u8f83\u597d\u7684\u7279\u5f81\u7ef4\u6570\u3002\u65b0\u5efaClassify.py\uff0c\u8f93\u5165\u4ee5\u4e0b\u4ee3\u7801\uff1a<\/p>\n
\n# coding:utf-8\nimport sys\nfrom numpy import *\nimport os\nfrom sklearn import feature_extraction\nfrom sklearn.feature_extraction.text import TfidfTransformer\nfrom sklearn.feature_extraction.text import CountVectorizer\nfrom sklearn.feature_selection import SelectKBest\nfrom sklearn.feature_selection import chi2\nfrom sklearn import svm\nimport matplotlib.pyplot as plt   #\u5bfc\u5165pyplot\u5b50\u5e93\n\nreload(sys)\nsys.setdefaultencoding('utf8')\n\ndef LoadData(l=1000):#\u8fd4\u56de\u8bad\u7ec3\u96c6\uff0c\u6807\u7b7e\uff0c\u8bcd\u9891\u77e9\u9635\uff0c\u8bcd\u888b\n    originFile = open("newtrain.csv","r")#\u521d\u59cb\u7684\u8bad\u7ec3\u96c6\u6587\u4ef6\n    line = originFile.readline()\n    q = 0\n    corpus=[]\n    target=[]\n    while line and q < l: \n        tmpdatas = line.split(' ')\n        datas = " ".join(tmpdatas[5:])\n        corpus.append(datas)\n        target.append(tmpdatas[2])\n        q += 1\n        line = originFile.readline()\n    vectorizer = CountVectorizer()#\u8be5\u7c7b\u4f1a\u5c06\u6587\u672c\u4e2d\u7684\u8bcd\u8bed\u8f6c\u6362\u4e3a\u8bcd\u9891\u77e9\u9635\uff0c\u77e9\u9635\u5143\u7d20a[i][j]\u8868\u793aj\u8bcd\u5728i\u7c7b\u6587\u672c\u4e0b\u7684\u8bcd\u9891 \n    wordcount = vectorizer.fit_transform(corpus)#\u5c06\u6587\u672c\u8f6c\u4e3a\u8bcd\u9891\u77e9\u9635\n    wordlist = vectorizer.get_feature_names()#\u83b7\u53d6\u8bcd\u888b\u6a21\u578b\u4e2d\u7684\u6240\u6709\u8bcd\u8bed\n    originfile.close()\n    return corpus,target,wordcount,wordlist\n\ndef chi(wordcount,target,k=200):#\u8fd4\u56de\u88ab\u9009\u7684k\u4e2a\u7279\u5f81\u6784\u6210\u7684\u7528\u6237\u77e9\u9635" \n    res = SelectKBest(chi2, k).fit_transform(wordcount.toarray(), target)\n    return res\n\ndef test(l=1000,k=200,j=200,wordCount=None,target=None):#\u5f97\u5230\u9884\u6d4b\u6b63\u786e\u6570\u4e0e\u9519\u8bef\u6570\n    res = chi(wordCount,target,j)\n    X = res[:-k]\n    y = target[:-k]\n    clf = svm.SVC()\n    clf.fit(X, y)\n    result = clf.predict(res[-k:])\n\n    yes = 0\n    no = 0\n    for i in range(len(result)):\n        if self.target[i + self.allNum - self.testNum] == result[i]:\n            yes += 1\n        else:\n            no += 1\n\n    return yes,no\n<\/code><\/pre>\n
\u8fd9\u6837\u6211\u4eec\u5c31\u5b8c\u6210\u4e86\u5bf9200\u4e2a\u7528\u6237\u7684\u6027\u522b\u5206\u7c7b\uff0c\u5e76\u6839\u636e\u7279\u5f81\u7ef4\u6570\u7684\u9009\u62e9\u60c5\u51b5\u5c06\u6b63\u786e\u7387\u53cd\u6620\u5728\u56fe\u4e2d\uff1a
\n\"\"<\/a>
\n\u5f53\u7ef4\u5ea6\u9009\u5728450-500\u65f6\uff0c\u6b63\u786e\u7387\u57fa\u672c\u7a33\u5b9a\u572874%\u3002<\/p>\n
\u901a\u8fc7\u4ee5\u4e0a\u8fd9\u4e9b\u6b65\u9aa4\uff0c\u6211\u4eec\u5c31\u5b8c\u6210\u4e86\u57fa\u672c\u7684\u6587\u672c\u5206\u7c7b\u5de5\u4f5c\uff0c\u4f46\u662f\u8fd9\u4ec5\u4ec5\u662f\u4e2a\u5f00\u59cb\u3002\u4ecd\u6709\u5f88\u591a\u95ee\u9898\u9700\u8981\u89e3\u51b3\uff1a
\n1.\u6709\u591a\u4e2a\u5c5e\u6027\u7684\u5e74\u9f84\u3001\u5b66\u5386\u5982\u4f55\u5206\u7c7b\uff1f
\n2.\u5982\u4f55\u5bf910w\u7528\u6237\u505a\u6587\u672c\u5206\u7c7b\uff1f
\n3.\u5982\u4f55\u63d0\u9ad8\u5206\u7c7b\u7684\u51c6\u786e\u6027\uff1f
\n4.\u5982\u4f55\u6574\u7406\u5df2\u5b8c\u6210\u7684\u4ee3\u7801\uff1f
\n\u8fd9\u4e9b\u95ee\u9898\u5c06\u5728\u4e4b\u540e\u7684\u6b65\u9aa4\u5c1d\u8bd5\u89e3\u51b3<\/p>\n
\n
4.\u5206\u7c7b<\/h3>\n
\u5176\u5b9e\u8fd9\u90e8\u5206\u5185\u5bb9\u548c\u4e0a\u9762\u7684\u7279\u5f81\u9009\u62e9\u5408\u5e76\u5199\u5230\u4e00\u4e2a\u51fd\u6570\uff0c\u901a\u8fc7skl\u7684svm\u51fd\u6570\uff0c\u6211\u4eec\u5c31\u53ef\u4ee5\u76f4\u63a5\u4f7f\u7528svm\u5206\u7c7b\u5668\u800c\u4e0d\u7528\u81ea\u5df1\u624b\u5de5\u7f16\u5199\u3002\u6240\u4ee5\u8fd9\u90e8\u5206\u5c06\u4e0d\u518d\u63cf\u8ff0\u5176\u4ed6\u5de5\u4f5c\u3002\u53ea\u6709\u4e00\u70b9\uff0c\u5bf9\u4e4b\u524d\u5df2\u5199\u597d\u7684\u4ee3\u7801\u8fdb\u884c\u91cd\u6784\uff0c\u628a\u4ee5\u4e0a\u5199\u597d\u7684py\u4ee3\u7801\u6574\u5408\u5230\u4e00\u4e2a\u4e2a\u51fd\u6570\u4e2d\uff0c\u5e76\u65b0\u5efamain.py\u5199\u5165\u8c03\u7528\u65b9\u6cd5\u3002
\n\u65b0\u5efamain.py\u5e76\u5199\u5165\u5982\u4e0b\u4ee3\u7801\uff1a<\/p>\n
\n#encoding=utf-8\nimport PrepareWork as p0\nimport DelWord as p1\nimport CountInfo as p2\nimport Classify as p3\n#p0.InitOriginFile("user_tag_query.10W.TRAIN","train.csv")\n#p1.DeleteStopWord("train.csv","newTrain.csv")\n#p2.CountUserRate("newTrain.csv","info.csv")\n<\/code><\/pre>\n
\u8fd9\u6837\u6211\u4eec\u628a\u4e4b\u524d\u7684\u4ee3\u7801\u90fd\u5c01\u88c5\u5728\u51fd\u6570\u4e2d\uff0c\u5e76\u53ef\u4ee5\u5728main.py\u8c03\u7528\u3002\u56e0\u4e3a\u8fd9\u4e9b\u6b65\u9aa4\u7684\u7ed3\u679c\u662f\u4ea7\u751fnewTrain.csv\u548cinfo.csv\uff0c\u800c\u6211\u4eec\u5df2\u7ecf\u751f\u6210\u4e86\u8fd9\u4e9b\u6587\u4ef6\uff0c\u6240\u4ee5\u5728main.py\u5c06\u4ed6\u4eec\u6ce8\u91ca\u6389\u3002<\/p>\n
\u53e6\u5916\uff0cClassify.py\u6bcf\u4e00\u6b21\u90fd\u6267\u884c\u4e86\u8bfb\u5199\u6587\u4ef6\u64cd\u4f5c\uff0c\u5b9e\u9645\u4e0a\u53ea\u9700\u8981\u8bfb\u53d6\u4e00\u6b21\uff0c\u5c06\u90a3\u4e9b\u8bad\u7ec3\u6837\u672c\u4fdd\u5b58\u4e0b\u6765\u5c31\u53ef\u4ee5\u4e86\u3002\u6240\u4ee5\u6211\u4eec\u5c06\u5176\u5c01\u88c5\u5728\u7c7b\u4e2d\uff0c\u907f\u514d\u91cd\u590d\u64cd\u4f5c\u3002\u91cd\u5199\u540e\u7684\u4ee3\u7801\u4e0d\u5728\u8fd9\u91cc\u5217\u51fa\uff0c\u53ef\u4ee5\u8bbf\u95ee\u6211\u7684github<\/a>\u83b7\u53d6\u4ee3\u7801<\/p>\n
\n
5.\u4f18\u5316<\/h3>\n
\u6709\u4e00\u53e5\u8bdd\u8bf4\u5f97\u5f88\u597d\uff0c\u5206\u7c7b\u7ed3\u679c\u7684\u6700\u5927\u51c6\u786e\u7a0b\u5ea6\u53d6\u51b3\u4e8e\u6570\u636e\u5904\u7406\u548c\u7279\u5f81\u9009\u62e9\uff0c\u800c\u4e3a\u4e86\u8fbe\u5230\u8fd9\u4e2a\u6700\u5927\u51c6\u786e\u5ea6\uff0c\u5c31\u9700\u8981\u8c03\u6574\u5206\u7c7b\u5668\u53c2\u6570\u3002
\n\u5728\u8fd9\u4e2a\u4f8b\u5b50\u4e2d\uff0c\u6211\u51b3\u5b9a\u4f7f\u7528\u6a21\u62df\u9000\u706b\u7b97\u6cd5\u5bf9SVM\u7684C\u53c2\u6570\u4ee5\u53ca\u7279\u5f81\u7ef4\u6570K\u53c2\u6570\u4f18\u5316\u3002\u8ba9\u6211\u4eec\u56de\u987e\u4e00\u4e0b\u4e4b\u524d\u7684\u5de5\u4f5c\u3002<\/p>\n
1.\u5207\u5206\u6240\u6709\u7528\u6237\u7684\u6240\u6709\u8bed\u53e5\u6210\u4e3a\u5355\u8bcd\uff0c\u5e76\u5c06\u8fd9\u4e9b\u5355\u8bcd\u52a0\u5165\u8bcd\u888b
\n2.\u4f7f\u7528\u5361\u65b9\u68c0\u9a8c\u9009\u51faK\u4e2a\u6700\u5177\u6709\u8868\u73b0\u529b\u7684\u5355\u8bcd\u8868\u793a\u6bcf\u4e00\u4e2a\u7528\u6237
\n3.\u5c06\u7528\u6237\u4e2dK\u4e2a\u5355\u8bcd\u7684\u51fa\u73b0\u6b21\u6570\u7ec4\u6210\u5411\u91cf\uff0c\u8f93\u5165\u5230SVM\u5206\u7c7b\u5668
\n4.\u4f7f\u7528SVM\u9884\u6d4b\u6d4b\u8bd5\u96c6\u7684\u6027\u522b<\/p>\n
K\u503c\u6211\u4eec\u5df2\u7ecf\u4e86\u89e3\u4e86\uff0c\u5c31\u662f\u6307\u9009\u62e9\u7684\u7279\u5f81\u7ef4\u6570\uff0c\u90a3\u4e48C\u662f\u4ec0\u4e48\u3002
\nC\u662fSVM\u7684\u60e9\u7f5a\u53c2\u6570\uff0c\u89c4\u5b9a\u6838\u51fd\u6570\u4e0e\u8bad\u7ec3\u6570\u636e\u7684\u62df\u5408\u7a0b\u5ea6\uff0c\u53c2\u8003sklearn\u5173\u4e8esvm\u53c2\u6570\u8bf4\u660e\u6587\u6863<\/a><\/p>\n
K\u662f\u6574\u6570\uff0c\u6216\u8bb8\u8fd8\u80fd\u4f7f\u7528\u7a77\u4e3e\u6cd5\u6c42\u51fa\uff0c\u4f46\u662fC\u662f\u6d6e\u70b9\u6570\uff0c\u53c8\u5982\u4f55\u9009\u62e9\u597d\u7684C\u53c2\u6570\u63d0\u9ad8\u5206\u7c7b\u7684\u51c6\u786e\u7a0b\u5ea6\u5462\uff1f
\nC\u7684\u53d6\u503c\u95ee\u9898\u5176\u5b9e\u7c7b\u4f3c\u4e8e\u51fd\u6570\u6c42\u6781\u503c\u3002\u8bd5\u60f3\u6211\u4eec\u5047\u8bbeK\u662f\u56fa\u5b9a\u7684\uff0c\u53ea\u662f\u4e3a\u4e86\u5bfb\u627e\u4e00\u4e2aC\u4f7f\u5f97SVM\u9884\u6d4b\u7ed3\u679ccorrect\u63d0\u5347\uff0c\u628asvm\u770b\u505a\u4e00\u4e2a\u51fd\u6570\uff0cC\u7684\u53d6\u503c\u5c31\u8f6c\u6362\u4e3a\u6c42SVM\u51fd\u6570\u7684\u6700\u5927\u503c\u95ee\u9898\u3002\u6c42\u51fd\u6570\u6781\u503c\u65b9\u6cd5\u6709\u5f88\u591a\uff0c\u8fd9\u91cc\u6211\u4eec\u4f7f\u7528\u6a21\u62df\u9000\u706b\u7b97\u6cd5\uff08\u53ef\u53c2\u8003\u8fd9\u7bc7\u6587\u7ae0\u5927\u767d\u8bdd\u89e3\u6790\u6a21\u62df\u9000\u706b\u7b97\u6cd5<\/a>\uff09<\/p>\n
\u6a21\u62df\u9000\u706b\u601d\u8def\u5f88\u7b80\u5355\uff0c\u4f8b\u5982\u6c42\u51fd\u6570y=x^5-x^4+x^2\u6781\u503c\uff0c\u6b65\u9aa4\u4e3a\uff1a<\/p>\n
1.\u786e\u5b9a\u6700\u5927\u6b65\u957fstep(x\u6bcf\u6b21\u524d\u8fdb\u6216\u540e\u9000\u7684\u6700\u5927\u503c)\uff0cx\u53d6\u503c\u8303\u56f4L(x\u5728\u533a\u95f4L\u5185\u53d6\u503c)\uff0c\u53d6\u5386\u53f2\u6700\u9ad8\u503c\u65f6X\u7684\u53d6\u503cBestX\uff0cy\u7684\u53d6\u503cBestScore\u53d6\u6700\u9ad8\u6e29\u5ea6T\uff0c\u53d6\u6700\u4f4e\u6e29\u5ea6minT
\n2.\u968f\u673a\u9009\u4e00\u4e2ax\u521d\u59cb\u503c\uff0c
\n3.x\u6309\u7167\u6b65\u957fstep\u524d\u8fdb(x+step)\u6216\u540e\u9000(x-step)\u5f97\u5230x1
\n4.\u6bd4\u8f83\u524d\u8fdb\u6216\u540e\u9000\u65f6\u7684y\u503c\u662f\u5426\u589e\u5927\uff0c
\n\u589e\u5927\uff1a\u66f4\u53d8\u540e\u7684x1\u66ff\u6362x\uff0c\u540c\u65f6\u66ff\u6362BestScore
\n\u51cf\u5c0f\uff1a\u4e00\u5b9a\u6982\u7387\u6267\u884c\u2018\u589e\u5927\u2019\u7684\u76f8\u540c\u64cd\u4f5c\uff0c\u4e00\u5b9a\u6982\u7387\u62d2\u7edd\u6539\u53d8x
\n5.T\u51cf\u5c0f\uff0c\u5982\u679cT\u5927\u4e8eminT\u5219\u91cd\u590d3\u64cd\u4f5c\uff0c\u5426\u5219\u505c\u6b62
\n6.\u8f93\u51faBestX\u548cBestScore<\/p>\n
\u5f53\u7136\u6a21\u62df\u9000\u706b\u4e5f\u53ef\u4ee5\u6c42\u5f97\u591a\u5143\u51fd\u6570\u6781\u503c\uff0c\u6211\u4eec\u8fd0\u7528\u4ee5\u4e0a\u601d\u60f3\uff0c\u4f7f\u7528\u6a21\u62df\u9000\u706b\u7b97\u6cd5\u786e\u5b9aK\u548cC\u7684\u503c\uff0c\u5728Classify\u4e2d\u6dfb\u52a0\u4ee3\u7801\uff1a<\/p>\n
\ndef Annealing(self,minTemp=0,minScore=100,KSTEP=5,CSTEP=2):\n    #\u641c\u7d22\u7684\u6700\u5927\u8303\u56f4\n    KMAX,CMAX=2000,1000           #\u51b7\u5374\u8868\u53c2\u6570\n    MarkovLength = 10000;         #\u9a6c\u53ef\u592b\u94fe\u957f\u5ea6\n    DecayScale = 0.98             #\u8870\u51cf\u53c2\u6570\n    Temperature = 100             #\u521d\u59cb\u6e29\u5ea6\n    PreK,NextK = 0,0                #prior and next value of x\n    PreC,NextC = 0.0,0.0            #prior and next value of y\n    BestK,BestC = 0,0.0             #\u6700\u7ec8\u89e3\n    PreScore,BestScore,Score = 0.0,0.0,0.0#\u5386\u53f2\u6210\u7ee9\uff0c\u6700\u597d\u6210\u7ee9\uff0c\u5f53\u524d\u6210\u7ee9\n    maxValue = 1000000\n\n    #\u968f\u673a\u9009\u70b9\n    PreK = 1#K\u503c\n    PreC = 0#C\u503c\n    PreBestK = BestK = PreK\n    PreBestC = BestC = PreC\n    \n    yes,no = 0,0\n    i = 0\n    #\u6e29\u5ea6\u8fc7\u4f4e\u6216\u5206\u6570\u8fbe\u5230\u8981\u6c42\u65f6\u505c\u6b62\n    while Temperature > minTemp and BestScore < minscore: \n        kstep,cstep = maxValue,maxValue\n        nextk,nextc = PreK + Kstep,PreC + Cstep #\u5982\u679c\u4e0b\u4e00\u6b65\u8d8a\u754c\u4e86\uff0c\u91cd\u65b0\u5b9a\u6b65\u957f\n        while nextk > KMAX or NextK <= 0 or NextC > CMAX or NextC < 0:\n            Kstep = r.randint(0,KSTEP) - KSTEP \/ 2\n            Cstep = CSTEP*(r.random() - 0.5)\n            NextK,NextC = PreK + Kstep,PreC + Cstep\n        yes,no,Score = self.CalScore(NextK,NextC)\n        print "i:",i,"k:",NextK,"c:",NextC,"correct:",Score,"pre:",PreScore,"T:",Temperature,\n        if Score > PreScore :#\u6bd4\u4e0a\u4e00\u4e2a\u89e3\u66f4\u597d\uff0c\u63a5\u53d7\n            if Score > BestScore:#\u6bd4\u6700\u597d\u89e3\u8fd8\u597d\uff0c\u66ff\u6362\n                BestScore = Score\n                BestK = NextK\n                BestC = NextC\n            PreK = NextK\n            PreC = NextC\n            PreScore = Score\n            print "accept1"\n        else:#\u6bd4\u4e0a\u4e00\u4e2a\u89e3\u5dee\n            change=1000.0 * (Score - PreScore) \/ Temperature#\u964d\u4f4e\u7684\u6570\u503c(\u8f6c\u5316\u4e3a\u6b63\u6570)\n            #print change,m.exp(change),\n            if m.exp(change) > r.random():#\u6982\u7387\u63a5\u53d7\u8f83\u5dee\u89e3\n                PreK = NextK\n                PreC = NextC\n                PreScore = Score\n                print "accept2"\n            else:#\u4e0d\u63a5\u53d7\n                print "refuse"\n                pass\n            Temperature *= DecayScale\n            i += 1\n\n        return BestK,BestC,BestScore\n<\/code><\/pre>\n
\u6700\u540e\uff0c\u518d\u6539\u5199main.py\u4e2d\u7684\u4ee3\u7801\uff1a<\/p>\n
\n#encoding=utf-8\nimport PrepareWork as p0\nimport DelWord as p1\nimport CountInfo as p2\nimport Classify as p3\n#p0.InitOriginFile("user_tag_query.10W.TRAIN","train.csv")\n#p1.DeleteStopWord("train.csv","newTrain.csv")\n#p2.CountUserRate("newTrain.csv","info.csv")\n\ndef testAnn():\n    k,c,score = clf.Annealing(1,80,100,2)\n    print "Best is:","k=",k,"c=",c,"correct=",score<\/code>\n\ndef testClassify():\n    global yes,no\n    for i in range(1,2000):\n        yes,no = clf.test(i*10)\n        print \"k:\",i * 10,\"yes:\",yes,\"no:\",no,\"correct:\",yes * 1.0 \/ (yes + no)\n\nyes,no = 0,0\nallNum,testNum = 2000,400\nk,c,score = 1,1.0,0\nallNum,testNum = 2000,400\nclf=p3.Clf(allNum,testNum)\nclf.LoadData(allNum)\n\ntestAnn()\n<\/pre>\n
\u4ee5\u4e0a\u4ee3\u7801\u5c06\u8bad\u7ec3\u6837\u672c\u6269\u5145\u81f31600\uff0c\u6d4b\u8bd5\u6837\u672c\u6269\u5145\u4e3a400\uff0c\u4f7f\u7528\u6a21\u62df\u9000\u706b\u6c42\u5f97K\u548cC\u53c2\u6570\uff0c\u7ed3\u679c\u5982\u4e0b\uff1a
\n\"\"<\/a>
\n\u81f3\u6b64\uff0c\u6211\u4eec\u5c31\u5df2\u7ecf\u5b8c\u6210\u4e86\u8f83\u9ad8\u7ea7\u7684\u6587\u672c\u4e8c\u5206\u7c7b\u5de5\u4f5c\uff0c\u540c\u65f6\u8ba9\u4ee3\u7801\u7a0d\u5fae\u597d\u770b\u4e86\u4e00\u4e9b\u3002
\n\u8fd9\u91cc\u7684\u5b66\u4e60\u5c31\u544a\u4e00\u6bb5\u843d\uff0c\u5bf9\u4e8e\u5269\u4e0b\u7684\u95ee\u9898\uff08\u591a\u5206\u7c7b\u3001\u5927\u6587\u672c\u5206\u7c7b\u3001\u7cbe\u5ea6\u63d0\u5347\u7b49\uff09\u5c06\u5728\u4ee5\u540e\u7ee7\u7eed\u505a\u7814\u7a76\u548c\u63a2\u7d22\u3002<\/p>\n
Thanks~<\/p>\n
\u53c2\u8003\u6587\u732e\uff1a
\nScikit-learn\uff1aFeature extraction\u6587\u672c\u7279\u5f81\u63d0\u53d6<\/a>
\n\u4f7f\u7528libsvm\u5b9e\u73b0\u6587\u672c\u5206\u7c7b<\/a>
\nTF-IDF\u4e0e\u4f59\u5f26\u76f8\u4f3c\u6027\u7684\u5e94\u7528<\/a>
\n\u6a21\u62df\u9000\u706b\u7b97\u6cd5\u4f8b\u5b50<\/a>
\nsklearn\u6587\u6863\u7d22\u5f15<\/a>
\n\u5361\u65b9\u68c0\u9a8c\u7528\u4e8e\u7279\u5f81\u9009\u62e9<\/a>
\ngithub\u53cb\u60c5\u94fe\u63a5<\/a><\/p>\n","protected":false},"excerpt":{"rendered":"
1,607 Views\u524d\u8a00 \u641c\u72d7\u7528\u6237\u753b\u50cf\u6316\u6398\u662f2016CCF\u5927\u6570\u636e\u7ade\u8d5b\u7684\u9898\u76ee\uff0c\u90a3\u65f6\u5019\u5bf9\u6570\u636e\u6316\u6398\u4e86\u89e3\u4e0d\u591a\uff0c\u518d\u52a0\u4e0a […]<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[26],"tags":[43,28,83,93],"class_list":["post-551","post","type-post","status-publish","format-standard","hentry","category-skl","tag-python","tag-sklearn","tag-83","tag-93"],"_links":{"self":[{"href":"http:\/\/www.mrtblog.cn\/index.php?rest_route=\/wp\/v2\/posts\/551","targetHints":{"allow":["GET"]}}],"collection":[{"href":"http:\/\/www.mrtblog.cn\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"http:\/\/www.mrtblog.cn\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"http:\/\/www.mrtblog.cn\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"http:\/\/www.mrtblog.cn\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=551"}],"version-history":[{"count":16,"href":"http:\/\/www.mrtblog.cn\/index.php?rest_route=\/wp\/v2\/posts\/551\/revisions"}],"predecessor-version":[{"id":1182,"href":"http:\/\/www.mrtblog.cn\/index.php?rest_route=\/wp\/v2\/posts\/551\/revisions\/1182"}],"wp:attachment":[{"href":"http:\/\/www.mrtblog.cn\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=551"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"http:\/\/www.mrtblog.cn\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=551"},{"taxonomy":"post_tag","embeddable":true,"href":"http:\/\/www.mrtblog.cn\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=551"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}