学習・推論パイプラインを構築する上で大切にしていること

kaggleやSIGNATEなどの分析コンペで使用する学習・推論パイプラインを構築する上で大切にしていること

Transcript

1. ֶशɾਪ࿦ύΠϓϥΠϯΛߏங͢Δ্Ͱ
   େ੾ʹ͍ͯ͠Δ͜ͱ $POOFIJUP
   *OD
   ໺ᖒ఩র
   
   ෼ੳίϯϖ-5ձ
2. ͜Μʹͪ͸ʂ 
3. ࠓ೔ൃද͢Δ͜ͱ͸Լهͱएׯॏෳ͠·͢  IUUQTTQFBLFSEFDLDPNUBLBQZEFUBGFOYJLPOQFOJPJUFUF[IFOHMJBOH HVBOMJOJQJCJTJUFJSVRVBOSFOMFJOJDIVBOFUBJYJBOHJ
4.  ΞδΣϯμ  ࣗݾ঺հ
    ύΠϓϥΠϯ͕ͳ͔ͬͨ࣌ͷπϥϛ
    େ੾ʹ͍ͯ͠Δ͜ͱ
   ‣ ֶशͷ࠶ݱੑ
   ‣ 1%$"ͷߴ଎Խ
   ‣ ҙࣝ͠ͳͯ͘΋؅ཧͰ͖Δ࢓૊Έ
    ࣮૷ৄࡉ
5. 
   ࣗݾ঺հ 
6.  
   ࣗݾ঺հ ໊લɿ໺ᖒ
   ఩রʢ/P[BXB
   5BLBOPCVʣ
   ॴଐɿίωώτגࣜձࣾ
   ɹɹɿ͔ͨͺ͍!UBLBQZ
   w ʙίωώτʹ.-ΤϯδχΞͱͯ͠+0*/
   w ػցֶशʢ/-1ɺਪનγεςϜʣΛϝΠϯʹ΍ΓͭͭΠϯϑϥʢ"84ʣ΋ษڧத
   w ,BHHMFͨ͠ΓɺϒϩάʢIUUQTXXXUBLBQZXPSLʣॻ͍ͨΓɺ໺ٿͨ͠Γɺ
   
   
   ϥʔϝϯ৯΂ͨΓ͍ͯ͠·͢
   w ΦεεϝͷυϥϜࣜચ୕ػ͋ͬͨΒڭ͍͑ͯͩ͘͞
7.  
   ࣗݾ঺հ ˞
   ʮӾཡ਺ʯʮར༻ऀ਺ʯ͸ϝσΟΞͱΞϓϦͷ߹ܭ஋ʢ೥݄݄ͷฏۉ஋ʣ
   ˞ʮϚϚ޲͚/PΞϓϦʯ͸೥݄Πϯςʔδௐ΂ɹௐࠪର৅ɿ೛৷தʙ̎ࡀ̌ϲ݄ͷࢠڙΛ࣋ͭঁੑ O  Λநग़
   ˞
   
   *OTUBHSBNͷϑΥϩϫʔ਺ɺ'BDFCPPLͷ͍͍Ͷ਺ɺ-*/&ͷͱ΋ͩͪ਺ͷ߹ܭ஋ ೥݄࣌఺
   ϚϚϦ
   ΞϓϦɾ8FC 4/4
   *OTUBHSBNɾ
   -*/&ɾ
   'BDFCPPL هࣄ ϚϚಉ࢜Ͱ೰ΈΛ૬ஊ͠߹͏2"ίϛϡχςΟΛத৺ʹ
   ϢʔβʔΛ֦େ͍ͯ͠·͢ ʮϚϚϦʯͰϢʔβʔಉ͕࢜
   ͲΜͲΜܨ͕͍ͬͯ·͢ ϚϚͷੜ׆ʹ໾ཱͭهࣄΛ
   ෯޿͍δϟϯϧͰ഑৴͍ͯ͠·͢ ϚϚ޲͚/P̍ΞϓϦʹબग़  ਓͷϚϚ͕બͿʮݱࡏ࢖͍ͬͯΔΞϓϦʯ
   ʹ ͯɺ߲໨
   ଞͷϚϚʹΦεεϝ͍ͨ͠ɺೝ஌౓ɺ
   ར༻཰ɺརศੑɺ޷ײ౓
   Ͱ̍ҐΛ֫ಘ͠·ͨ͠ هࣄ਺ 6,000 هࣄҎ্ ྦྷܭϑΝϯ਺ ໿ 85 ສਓ ˞ ݄ؒӾཡ਺ ໿ 1.5ԯճ ˞ ݄ؒར༻ऀ਺ ໿ 650ສਓ ˞ ˞ l೰Έz
   ͱ
   zڞײz
   Λ࣠ʹϚϚʹدΓఴ͍
   ΞϓϦɾ8FCɾ4/4ͱଟ֯తʹαʔϏεΛల։͍ͯ͠·͢
8.  
   ࣗݾ঺հ 0 450,000 900,000 1,350,000 1,800,000 2014/4 2014/5 2014/6 2014/7 2014/8 2014/9 2014/10 2014/11 2014/12 2015/1 2015/2 2015/3 2015/4 2015/5 2015/6 2015/7 2015/8 2015/9 2015/10 2015/11 2015/12 2016/1 2016/2 2016/3 2016/4 2016/5 2016/6 2016/7 2016/8 2016/9 2016/10 2016/11 2016/12 2017/1 2017/2 2017/3 2017/4 2017/5 2017/6 2017/7 2017/8 2017/9 2017/10 2017/11 2017/12 2018/1 2018/2 2018/3 2018/4 2018/5 2018/6 2018/7 2018/8 ݄ؒ౤ߘ਺ ໿ 150ສ݅ िʹ೔Ҏ্ىಈ͢Δ
   ΞΫςΟϒϢʔβʔ ໿ 50 ਓʹਓ 57$.
   ์ө ΞϓϦ૯%-਺
   ສ ਓʹਓ ਓʹਓ ਓʹਓ ਓʹਓ ˞ ˞
   ʮϚϚϦʯ಺ͷग़࢈༧ఆ೔Λઃఆͨ͠Ϣʔβʔ਺ͱɺްੜ࿑ಇলൃදʮਓޱಈଶ౷ܭʯͷग़ੜ਺͔Βࢉग़
   ˞
   िʹճҎ্ىಈ͢ΔϢʔβʔ ˞ ೥ʹग़࢈ͨ͠ϚϚͷʮਓʹਓʯ͕ϚϚϦΛར༻த
   ೔ຊ࠷େڃن໛ΛތΔϒϥϯυ΁ͱ੒௕͍ͯ͠·͢ ˞
9. 
   ύΠϓϥΠϯ͕ͳ͔ͬͨ࣌ͷπϥϛ 
10.  
    ύΠϓϥΠϯ͕ͳ͔ͬͨ࣌ͷπϥϛ ΊͬͪΌྑ͍είΞͷOPUFCPPL͕׬੒
    ˣ
    ͜ͷOPUFCPPLΛ%VQMJDBUFͯ͠ɺ΋ͬͱྑ͍Ϟσϧ࡞ͬͪΌ͏ͧʂ
11.  ҰํɺOPUFCPPLͷத਎͸ʜ
12.  
    ύΠϓϥΠϯ͕ͳ͔ͬͨ࣌ͷπϥϛ <> import numpy as np import pandas as pd OPUFCPPLͷத਎ ɾ
    ɾ
    ɾ <> hogehoge <> hogehoge <> hogehoge ɾ
    ɾ
    ɾ
13.  
    ύΠϓϥΠϯ͕ͳ͔ͬͨ࣌ͷπϥϛ <> import numpy as np import pandas as pd OPUFCPPLͷத਎ ɾ
    ɾ
    ɾ <> hogehoge <> hogehoge <> hogehoge ɾ
    ɾ
    ɾ ηϧͷ࣮ߦॱ͕ͪ͝Όͪ͝Ό
    ˣ
    ࠶ݱੑ͕ͳ͍
14.  
    ύΠϓϥΠϯ͕ͳ͔ͬͨ࣌ͷπϥϛ <> import numpy as np import pandas as pd OPUFCPPLͷத਎ ɾ
    ɾ
    ɾ <> submission.to_csv('submission.csv', index=False)
15.  
    ύΠϓϥΠϯ͕ͳ͔ͬͨ࣌ͷπϥϛ <> import numpy as np import pandas as pd OPUFCPPLͷத਎ ɾ
    ɾ
    ɾ <> submission.to_csv('submission.csv', index=False) ηϧ͕ଟ͘ɺಉ͡ܭࢉΛෳ਺࣮ߦ͠ͳ͚Ε͹ͳΒͳ͍
    ˣ
    1%$"͕஗͘ͳΓ͕ͪ
16.  
    ύΠϓϥΠϯ͕ͳ͔ͬͨ࣌ͷπϥϛ dOPUFCPPL-JHIU#(.@TDPSF@$PQZJQZOC
    dOPUFCPPL-JHIU#(.@TDPSF@$PQZJQZOC
    dOPUFCPPL-JHIU#(.@TDPSF@$PQZJQZOC
    dOPUFCPPL-JHIU#(.@TDPSF@$PQZJQZOC
    dOPUFCPPL-JHIU#(.@TDPSF@$PQZJQZOC
    dOPUFCPPL-JHIU#(.@TDPSF@$PQZJQZOC
    dOPUFCPPL-JHIU#(.@TDPSF@$PQZJQZOC
    dOPUFCPPL-JHIU#(.@TDPSF@$PQZJQZOC
    dOPUFCPPL-JHIU#(.@TDPSF@JQZOC
    dOPUFCPPL-JHIU#(.@TDPSF@JQZOC
    dOPUFCPPL-JHIU#(.@TDPSF@JQZOC
    dOPUFCPPL-JHIU#(.@TDPSF@JQZOC
    dOPUFCPPL-JHIU#(.@TDPSF@JQZOC
    dOPUFCPPL-JHIU#(.@TDPSF@JQZOC
    dOPUFCPPL-JHIU#(.@TDPSF@JQZOC
    dOPUFCPPL-JHIU#(.@TDPSF@JQZOC
    ʜʜʜʜʜ
17.  
    ύΠϓϥΠϯ͕ͳ͔ͬͨ࣌ͷπϥϛ dOPUFCPPL-JHIU#(.@TDPSF@$PQZJQZOC
    dOPUFCPPL-JHIU#(.@TDPSF@$PQZJQZOC
    dOPUFCPPL-JHIU#(.@TDPSF@$PQZJQZOC
    dOPUFCPPL-JHIU#(.@TDPSF@$PQZJQZOC
    dOPUFCPPL-JHIU#(.@TDPSF@$PQZJQZOC
    dOPUFCPPL-JHIU#(.@TDPSF@$PQZJQZOC
    dOPUFCPPL-JHIU#(.@TDPSF@$PQZJQZOC
    dOPUFCPPL-JHIU#(.@TDPSF@$PQZJQZOC
    dOPUFCPPL-JHIU#(.@TDPSF@JQZOC
    dOPUFCPPL-JHIU#(.@TDPSF@JQZOC
    dOPUFCPPL-JHIU#(.@TDPSF@JQZOC
    dOPUFCPPL-JHIU#(.@TDPSF@JQZOC
    dOPUFCPPL-JHIU#(.@TDPSF@JQZOC
    dOPUFCPPL-JHIU#(.@TDPSF@JQZOC
    dOPUFCPPL-JHIU#(.@TDPSF@JQZOC
    dOPUFCPPL-JHIU#(.@TDPSF@JQZOC
    ʜʜʜʜʜ OPUFCPPL͕ࡍݶͳ͘૿৩͠
    ؅ཧ͕ΊͪΌͪ͘Ό൥ࡶʹ
18.  ͜ΕΒͷπϥϛΛղফ͢΂͘
    ୤ɾOPUFCPPLʹνϟϨϯδ
    ˣ
    ͦΜͳதͰେ੾ʹ͍ͯ͠Δ͜ͱΛ͓఻͑͠·͢ ˞&%"ͳͲ͸OPUFCPPLΛ࢖༻
19.  ౔୆͸͜ͷຊͰ͢
20. 
    େ੾ʹ͍ͯ͠Δ͜ͱ 
21.  ͦͷ
    ʙֶशͷ࠶ݱੑʙ
22.  ͦͷ
    ʙֶशͷ࠶ݱੑʙ ϙΠϯτ͸̏ͭ
23.  
    େ੾ʹ͍ͯ͠Δ͜ͱ w ͲΜͳಛ௃ྔΛ࢖ͬͯ ʙֶशͷ࠶ݱੑʙ
24.  
    େ੾ʹ͍ͯ͠Δ͜ͱ w ͲΜͳಛ௃ྔΛ࢖ͬͯ
    w ͲΜͳύϥϝʔλʔΛ࢖ͬͯ ʙֶशͷ࠶ݱੑʙ
25.  
    େ੾ʹ͍ͯ͠Δ͜ͱ w ͲΜͳಛ௃ྔΛ࢖ͬͯ
    w ͲΜͳύϥϝʔλʔΛ࢖ͬͯ
    w ͲΜͳ$7Λ࢖ͬͯ
    ֶश͔ͨ͠Λอଘ͓ͯ͘͠ɻ ʙֶशͷ࠶ݱੑʙ
26.  
    େ੾ʹ͍ͯ͠Δ͜ͱ w ͲΜͳಛ௃ྔΛ࢖ͬͯ
    w ͲΜͳύϥϝʔλʔΛ࢖ͬͯ
    w ͲΜͳ$7Λ࢖ͬͯ
    ֶश͔ͨ͠Λอଘ͓ͯ͘͠ɻ ʙֶशͷ࠶ݱੑʙ 
    %0//"
    อଘଇ
27.  ͦͷ
    ʙߴ଎ͳ1%$"Λ໨ࢦͯ͠ʙ
28.  
    େ੾ʹ͍ͯ͠Δ͜ͱ w ͭͷQZUIPOεΫϦϓτͰ
    %0//"ΛঠѲ
     ࢖༻͢Δಛ௃ྔ
     ࢖༻͢Δύϥϝʔλ
     ࢖༻͢Δ$7
    ʙߴ଎ͳ1%$"Λ໨ࢦͯ͠ʙ
29.  
    େ੾ʹ͍ͯ͠Δ͜ͱ w ͭͷQZUIPOεΫϦϓτͰ
    %0//"ΛঠѲ
     ࢖༻͢Δಛ௃ྔ
     ࢖༻͢Δύϥϝʔλ
     ࢖༻͢Δ$7
    w GFBUVSF
    JNQPSUBODFͷݟ͑ΔԽ
     ࣍ͷ࣮ݧ΁ͷצॴΛ͔ͭΊΔΑ͏ʹ ʙߴ଎ͳ1%$"Λ໨ࢦͯ͠ʙ
30.  ͦͷ
    ʙҙࣝ͠ͳͯ͘΋؅ཧͰ͖Δ࢓૊Έʙ
31.  
    େ੾ʹ͍ͯ͠Δ͜ͱ ʙҙࣝ͠ͳͯ͘΋؅ཧͰ͖Δ࢓૊Έʙ w 
    %0///"
    
    Ϟσϧ΍ϩάϑΝΠϧΛɺ࣮ߦ͝ͱʹʮ೔࣌ʯ TV⒏YΛ͚ͭͨσΟϨΫτϦɾϑΝΠϧͰࣗಈ؅ཧ
     ࢖༻͢Δಛ௃ྔ
    ύϥϝʔλ
    $7
     ֶशͨ͠Ϟσϧ
     ֶशϩάϑΝΠϧ
     GFBUVSF
    JNQPSUBODF
    
32.  
    େ੾ʹ͍ͯ͠Δ͜ͱ ʙҙࣝ͠ͳͯ͘΋؅ཧͰ͖Δ࢓૊Έʙ w 
    %0///"
    
    Ϟσϧ΍ϩάϑΝΠϧΛɺ࣮ߦ͝ͱʹʮ೔࣌ʯ TV⒏YΛ͚ͭͨσΟϨΫτϦɾϑΝΠϧͰࣗಈ؅ཧ
     ࢖༻͢Δಛ௃ྔ
    ύϥϝʔλ
    $7
     ֶशͨ͠Ϟσϧ
     ֶशϩάϑΝΠϧ
     GFBUVSF
    JNQPSUBODF
    w 1VCMJDείΞΛσΟϨΫτϦͷ1SFpYʹ෇͚Δ͜ͱͰɺ 1VCMJDͱMPDBMͷείΞͷରԠΛ෼͔Γ΍͘͢ʢৄࡉ͸ޙ΄Ͳʣ
33.  
    େ੾ʹ͍ͯ͠Δ͜ͱ ʙҙࣝ͠ͳͯ͘΋؅ཧͰ͖Δ࢓૊Έʙ w 
    %0///"
    
    Ϟσϧ΍ϩάϑΝΠϧΛɺ࣮ߦ͝ͱʹʮ೔࣌ʯ TV⒏YΛ͚ͭͨσΟϨΫτϦɾϑΝΠϧͰࣗಈ؅ཧ
     ࢖༻͢Δಛ௃ྔ
    ύϥϝʔλ
    $7
     ֶशͨ͠Ϟσϧ
     ֶशϩάϑΝΠϧ
     GFBUVSF
    JNQPSUBODF
    w 1VCMJDείΞΛσΟϨΫτϦͷ1SFpYʹ෇͚Δ͜ͱͰɺ 1VCMJDͱMPDBMͷείΞͷରԠΛ෼͔Γ΍͘͢ʢৄࡉ͸ޙ΄Ͳʣ ৄࡉΛݟ͍͖ͯ·͢
34. 
    ࣮૷ৄࡉ 
35.  ྫʣIPHFQZ features = [ "age", "pclass", "family_size", "fare", "sibsp", "parch", "cabin" ] params_lgb = { 'boosting_type': 'gbdt', 'objective': 'fair', 'metric': 'fair', 'num_boost_round': 20000, 'early_stopping_rounds': 1000, 'verbose': 1000, 'random_state': 999 } cv = { 'method': 'KFold', 'n_splits': 5, 'random_state': 42, 'shuffle': True, } runner = Runner(run_name, ModelLGB, features, dataset.get('target'), params_lgb, cv, FEATURE_DIR_NAME, MODEL_DIR_NAME) runner.run_train_cv() # ֶश runner.run_predict_cv() # ਪ࿦ Submission.create_submission(run_name) # submit࡞੒
36.  ྫʣIPHFQZ features = [ "age", "pclass", "family_size", "fare", "sibsp", "parch", "cabin" ] params_lgb = { 'boosting_type': 'gbdt', 'objective': 'fair', 'metric': 'fair', 'num_boost_round': 20000, 'early_stopping_rounds': 1000, 'verbose': 1000, 'random_state': 999 } cv = { 'method': 'KFold', 'n_splits': 5, 'random_state': 42, 'shuffle': True, } runner = Runner(run_name, ModelLGB, features, dataset.get('target'), params_lgb, cv, FEATURE_DIR_NAME, MODEL_DIR_NAME) runner.run_train_cv() # ֶश runner.run_predict_cv() # ਪ࿦ Submission.create_submission(run_name) # submit࡞੒ جຊతʹ1%$"Λճ͢ͱ͖ʹ͍͡Δͷ͸
    ͜ͷ
    IPHFQZ
    ͷΈʹ͢Δ
37.  ྫʣIPHFQZ features = [ "age", "pclass", "family_size", "fare", "sibsp", "parch", "cabin" ] params_lgb = { 'boosting_type': 'gbdt', 'objective': 'fair', 'metric': 'fair', 'num_boost_round': 20000, 'early_stopping_rounds': 1000, 'verbose': 1000, 'random_state': 999 } cv = { 'method': 'KFold', 'n_splits': 5, 'random_state': 42, 'shuffle': True, } runner = Runner(run_name, ModelLGB, features, dataset.get('target'), params_lgb, cv, FEATURE_DIR_NAME, MODEL_DIR_NAME) runner.run_train_cv() # ֶश runner.run_predict_cv() # ਪ࿦ Submission.create_submission(run_name) # submit࡞੒ ֶशʹ࢖༻͢Δಛ௃ྔ
38.  ྫʣIPHFQZ features = [ "age", "pclass", "family_size", "fare", "sibsp", "parch", "cabin" ] params_lgb = { 'boosting_type': 'gbdt', 'objective': 'fair', 'metric': 'fair', 'num_boost_round': 20000, 'early_stopping_rounds': 1000, 'verbose': 1000, 'random_state': 999 } cv = { 'method': 'KFold', 'n_splits': 5, 'random_state': 42, 'shuffle': True, } runner = Runner(run_name, ModelLGB, features, dataset.get('target'), params_lgb, cv, FEATURE_DIR_NAME, MODEL_DIR_NAME) runner.run_train_cv() # ֶश runner.run_predict_cv() # ਪ࿦ Submission.create_submission(run_name) # submit࡞੒ ϋΠύʔύϥϝʔλ
    
39.  ྫʣIPHFQZ features = [ "age", "pclass", "family_size", "fare", "sibsp", "parch", "cabin" ] params_lgb = { 'boosting_type': 'gbdt', 'objective': 'fair', 'metric': 'fair', 'num_boost_round': 20000, 'early_stopping_rounds': 1000, 'verbose': 1000, 'random_state': 999 } cv = { 'method': 'KFold', 'n_splits': 5, 'random_state': 42, 'shuffle': True, } runner = Runner(run_name, ModelLGB, features, dataset.get('target'), params_lgb, cv, FEATURE_DIR_NAME, MODEL_DIR_NAME) runner.run_train_cv() # ֶश runner.run_predict_cv() # ਪ࿦ Submission.create_submission(run_name) # submit࡞੒ $7ͷઃఆ
40.  ྫʣIPHFQZ features = [ "age", "pclass", "family_size", "fare", "sibsp", "parch", "cabin" ] params_lgb = { 'boosting_type': 'gbdt', 'objective': 'fair', 'metric': 'fair', 'num_boost_round': 20000, 'early_stopping_rounds': 1000, 'verbose': 1000, 'random_state': 999 } cv = { 'method': 'KFold', 'n_splits': 5, 'random_state': 42, 'shuffle': True, } runner = Runner(run_name, ModelLGB, features, dataset.get('target'), params_lgb, cv, FEATURE_DIR_NAME, MODEL_DIR_NAME) runner.run_train_cv() # ֶश runner.run_predict_cv() # ਪ࿦ Submission.create_submission(run_name) # submit࡞੒ ֶशɾਪ࿦ɾTVC࡞੒
41.  ྫʣIPHFQZ features = [ "age", "pclass", "family_size", "fare", "sibsp", "parch", "cabin" ] params_lgb = { 'boosting_type': 'gbdt', 'objective': 'fair', 'metric': 'fair', 'num_boost_round': 20000, 'early_stopping_rounds': 1000, 'verbose': 1000, 'random_state': 999 } cv = { 'method': 'KFold', 'n_splits': 5, 'random_state': 42, 'shuffle': True, } runner = Runner(run_name, ModelLGB, features, dataset.get('target'), params_lgb, cv, FEATURE_DIR_NAME, MODEL_DIR_NAME) runner.run_train_cv() # ֶश runner.run_predict_cv() # ਪ࿦ Submission.create_submission(run_name) # submit࡞੒ ֶशɾਪ࿦ɾTVC࡞੒ w ࠶ݱੑΛ୲อ͢Δ޻෉
    w ߴ଎ͳ1$%"Λճͨ͢Ίͷ޻෉
    w ҙࣝ͠ͳͯ͘΋ॾʑ͕؅ཧͰ͖Δ޻෉
    ʹ͍͓ͭͯ࿩͠͠·͢
42.  ʙֶशͷ࠶ݱੑΛ୲อ͢ΔͨΊͷ޻෉ʙ
43.  
    ࣮૷ৄࡉ IPHFQZ ʙֶशͷ࠶ݱੑʙ features = [ "age", "pclass", "family_size", "fare", "sibsp", "parch", "cabin" ] params_lgb = { 'boosting_type': 'gbdt', 'objective': 'fair', 'metric': 'fair', 'num_boost_round': 20000, 'early_stopping_rounds': 1000, 'verbose': 1000, 'random_state': 999 } cv = { 'method': 'KFold', 'n_splits': 5, 'random_state': 42, 'shuffle': True, }
44.  
    ࣮૷ৄࡉ IPHFQZ ʙֶशͷ࠶ݱੑʙ features = [ "age", "pclass", "family_size", "fare", "sibsp", "parch", "cabin" ] params_lgb = { 'boosting_type': 'gbdt', 'objective': 'fair', 'metric': 'fair', 'num_boost_round': 20000, 'early_stopping_rounds': 1000, 'verbose': 1000, 'random_state': 999 } cv = { 'method': 'KFold', 'n_splits': 5, 'random_state': 42, 'shuffle': True, } 
    QZUIPO
    IPHFQZ
45.  
    ࣮૷ৄࡉ IPHFQZ ʙֶशͷ࠶ݱੑʙ features = [ "age", "pclass", "family_size", "fare", "sibsp", "parch", "cabin" ] params_lgb = { 'boosting_type': 'gbdt', 'objective': 'fair', 'metric': 'fair', 'num_boost_round': 20000, 'early_stopping_rounds': 1000, 'verbose': 1000, 'random_state': 999 } cv = { 'method': 'KFold', 'n_splits': 5, 'random_state': 42, 'shuffle': True, } { "use_features": [ "age", "pclass", "family_size", "fare", "sibsp", "parch", "cabin" ], "model_params": { "boosting_type": "gbdt", "objective": "fair", "metric": "fair", "num_boost_round": 20000, "early_stopping_rounds": 1000, "verbose": 1000, "random_state": 999 }, "cv": { "method": "KFold", "n_splits": 5, "random_state": 42, "shuffle": true }, "dataset": { "run_name": "lgb_1128_2003", "feature_directory": "../data/features/remove_outlier/", "target": "salary" } } IPHF@QBSBNKTPO 
    QZUIPO
    IPHFQZ
46.  
    ࣮૷ৄࡉ IPHFQZ ʙֶशͷ࠶ݱੑʙ features = [ "age", "pclass", "family_size", "fare", "sibsp", "parch", "cabin" ] params_lgb = { 'boosting_type': 'gbdt', 'objective': 'fair', 'metric': 'fair', 'num_boost_round': 20000, 'early_stopping_rounds': 1000, 'verbose': 1000, 'random_state': 999 } cv = { 'method': 'KFold', 'n_splits': 5, 'random_state': 42, 'shuffle': True, } { "use_features": [ "age", "pclass", "family_size", "fare", "sibsp", "parch", "cabin" ], "model_params": { "boosting_type": "gbdt", "objective": "fair", "metric": "fair", "num_boost_round": 20000, "early_stopping_rounds": 1000, "verbose": 1000, "random_state": 999 }, "cv": { "method": "KFold", "n_splits": 5, "random_state": 42, "shuffle": true }, "dataset": { "run_name": "lgb_1128_2003", "feature_directory": "../data/features/remove_outlier/", "target": "salary" } } IPHF@QBSBNKTPO 
    QZUIPO
    IPHFQZ IPHFQZΛ࣮ߦ͢Δ͜ͱʹΑΓ
    ࣗಈతʹKTPOϑΝΠϧ͕ੜ੒͞Ε
    ࢖༻ͨ͠ಛ௃ྔɾύϥϝʔλʔͳͲ͕શͯอଘ͞ΕΔ
47.  
    ࣮૷ৄࡉ IPHFQZ ʙֶशͷ࠶ݱੑʙ features = [ "age", "pclass", "family_size", "fare", "sibsp", "parch", "cabin" ] params_lgb = { 'boosting_type': 'gbdt', 'objective': 'fair', 'metric': 'fair', 'num_boost_round': 20000, 'early_stopping_rounds': 1000, 'verbose': 1000, 'random_state': 999 } cv = { 'method': 'KFold', 'n_splits': 5, 'random_state': 42, 'shuffle': True, } { "use_features": [ "age", "pclass", "family_size", "fare", "sibsp", "parch", "cabin" ], "model_params": { "boosting_type": "gbdt", "objective": "fair", "metric": "fair", "num_boost_round": 20000, "early_stopping_rounds": 1000, "verbose": 1000, "random_state": 999 }, "cv": { "method": "KFold", "n_splits": 5, "random_state": 42, "shuffle": true }, "dataset": { "run_name": "lgb_1128_2003", "feature_directory": "../data/features/remove_outlier/", "target": "salary" } } IPHF@QBSBNKTPO 
    QZUIPO
    IPHFQZ ͜Εͧ࠶ݱੑ
48.  ʙߴ଎ͳ1%$"Λճͨ͢Ίͷ޻෉ʙ
49.  
    ࣮૷ৄࡉ IPHFQZ features = [ "age", "pclass", "family_size", "fare", "sibsp", "parch", "cabin" ] params_lgb = { 'boosting_type': 'gbdt', 'objective': 'fair', 'metric': 'fair', 'num_boost_round': 20000, 'early_stopping_rounds': 1000, 'verbose': 1000, 'random_state': 999 } cv = { 'method': 'KFold', 'n_splits': 5, 'random_state': 42, 'shuffle': True, } ʙߴ଎ͳ1%$"Λ໨ࢦͯ͠ʙ
50.  
    ࣮૷ৄࡉ IPHFQZ features = [ "age", "pclass", "family_size", "fare", "sibsp", "parch", "cabin" ] params_lgb = { 'boosting_type': 'gbdt', 'objective': 'fair', 'metric': 'fair', 'num_boost_round': 20000, 'early_stopping_rounds': 1000, 'verbose': 1000, 'random_state': 999 } cv = { 'method': 'KFold', 'n_splits': 5, 'random_state': 42, 'shuffle': True, } ʙߴ଎ͳ1%$"Λ໨ࢦͯ͠ʙ ಛ௃ྔΛݮΒֶͯ͠श͍ͤͨ͞
51.  
    ࣮૷ৄࡉ IPHFQZ features = [ "age", "pclass", "family_size", "fare", "sibsp", "parch", "cabin" ] params_lgb = { 'boosting_type': 'gbdt', 'objective': 'fair', 'metric': 'fair', 'num_boost_round': 20000, 'early_stopping_rounds': 1000, 'verbose': 1000, 'random_state': 999 } cv = { 'method': 'KFold', 'n_splits': 5, 'random_state': 42, 'shuffle': True, } ʙߴ଎ͳ1%$"Λ໨ࢦͯ͠ʙ features = [ "age", "pclass", # "family_size", "fare", "sibsp", "parch", "cabin" ] params_lgb = { 'boosting_type': 'gbdt', 'objective': 'fair', 'metric': 'fair', 'num_boost_round': 20000, 'early_stopping_rounds': 1000, 'verbose': 1000, 'random_state': 999 } cv = { 'method': 'KFold', 'n_splits': 5, 'random_state': 42, 'shuffle': True, } IPHFQZ ର৅ͷಛ௃ྔΛίϝϯτΞ΢τ
52.  
    ࣮૷ৄࡉ IPHFQZ features = [ "age", "pclass", "family_size", "fare", "sibsp", "parch", "cabin" ] params_lgb = { 'boosting_type': 'gbdt', 'objective': 'fair', 'metric': 'fair', 'num_boost_round': 20000, 'early_stopping_rounds': 1000, 'verbose': 1000, 'random_state': 999 } cv = { 'method': 'KFold', 'n_splits': 5, 'random_state': 42, 'shuffle': True, } ʙߴ଎ͳ1%$"Λ໨ࢦͯ͠ʙ features = [ "age", "pclass", # "family_size", "fare", "sibsp", "parch", "cabin" ] params_lgb = { 'boosting_type': 'gbdt', 'objective': 'fair', 'metric': 'fair', 'num_boost_round': 20000, 'early_stopping_rounds': 1000, 'verbose': 1000, 'random_state': 999 } cv = { 'method': 'KFold', 'n_splits': 5, 'random_state': 42, 'shuffle': True, } IPHFQZ 
    QZUIPO
    IPHFQZ ର৅ͷಛ௃ྔΛίϝϯτΞ΢τ
53.  
    ࣮૷ৄࡉ IPHFQZ features = [ "age", "pclass", "family_size", "fare", "sibsp", "parch", "cabin" ] params_lgb = { 'boosting_type': 'gbdt', 'objective': 'fair', 'metric': 'fair', 'num_boost_round': 20000, 'early_stopping_rounds': 1000, 'verbose': 1000, 'random_state': 999 } cv = { 'method': 'KFold', 'n_splits': 5, 'random_state': 42, 'shuffle': True, } ʙߴ଎ͳ1%$"Λ໨ࢦͯ͠ʙ features = [ "age", "pclass", # "family_size", "fare", "sibsp", "parch", "cabin" ] params_lgb = { 'boosting_type': 'gbdt', 'objective': 'fair', 'metric': 'fair', 'num_boost_round': 20000, 'early_stopping_rounds': 1000, 'verbose': 1000, 'random_state': 999 } cv = { 'method': 'KFold', 'n_splits': 5, 'random_state': 42, 'shuffle': True, } IPHFQZ 
    QZUIPO
    IPHFQZ ର৅ͷಛ௃ྔΛίϝϯτΞ΢τ ؆୯
54.  
    ࣮૷ৄࡉ IPHFQZ features = [ "age", "pclass", "family_size", "fare", "sibsp", "parch", "cabin" ] params_lgb = { 'boosting_type': 'gbdt', 'objective': 'fair', 'metric': 'fair', 'num_boost_round': 20000, 'early_stopping_rounds': 1000, 'verbose': 1000, 'random_state': 999 } cv = { 'method': 'KFold', 'n_splits': 5, 'random_state': 42, 'shuffle': True, } ʙߴ଎ͳ1%$"Λ໨ࢦͯ͠ʙ
55.  
    ࣮૷ৄࡉ IPHFQZ features = [ "age", "pclass", "family_size", "fare", "sibsp", "parch", "cabin" ] params_lgb = { 'boosting_type': 'gbdt', 'objective': 'fair', 'metric': 'fair', 'num_boost_round': 20000, 'early_stopping_rounds': 1000, 'verbose': 1000, 'random_state': 999 } cv = { 'method': 'KFold', 'n_splits': 5, 'random_state': 42, 'shuffle': True, } ʙߴ଎ͳ1%$"Λ໨ࢦͯ͠ʙ $7มֶ͑ͯश͍ͤͨ͞
56.  
    ࣮૷ৄࡉ IPHFQZ features = [ "age", "pclass", "family_size", "fare", "sibsp", "parch", "cabin" ] params_lgb = { 'boosting_type': 'gbdt', 'objective': 'fair', 'metric': 'fair', 'num_boost_round': 20000, 'early_stopping_rounds': 1000, 'verbose': 1000, 'random_state': 999 } cv = { 'method': 'KFold', 'n_splits': 5, 'random_state': 42, 'shuffle': True, } ʙߴ଎ͳ1%$"Λ໨ࢦͯ͠ʙ features = [ "age", "pclass", "family_size", "fare", "sibsp", "parch", "cabin" ] params_lgb = { 'boosting_type': 'gbdt', 'objective': 'fair', 'metric': 'fair', 'num_boost_round': 20000, 'early_stopping_rounds': 1000, 'verbose': 1000, 'random_state': 999 } cv = { 'method': 'GroupKFold', 'n_splits': 5, 'random_state': 42, 'shuffle': True, 'cv_target':'user' } IPHFQZ DWͷهࡌΛมߋ
57.  
    ࣮૷ৄࡉ IPHFQZ features = [ "age", "pclass", "family_size", "fare", "sibsp", "parch", "cabin" ] params_lgb = { 'boosting_type': 'gbdt', 'objective': 'fair', 'metric': 'fair', 'num_boost_round': 20000, 'early_stopping_rounds': 1000, 'verbose': 1000, 'random_state': 999 } cv = { 'method': 'KFold', 'n_splits': 5, 'random_state': 42, 'shuffle': True, } ʙߴ଎ͳ1%$"Λ໨ࢦͯ͠ʙ features = [ "age", "pclass", "family_size", "fare", "sibsp", "parch", "cabin" ] params_lgb = { 'boosting_type': 'gbdt', 'objective': 'fair', 'metric': 'fair', 'num_boost_round': 20000, 'early_stopping_rounds': 1000, 'verbose': 1000, 'random_state': 999 } cv = { 'method': 'GroupKFold', 'n_splits': 5, 'random_state': 42, 'shuffle': True, 'cv_target':'user' } IPHFQZ DWͷهࡌΛมߋ 
    QZUIPO
    IPHFQZ
58.  
    ࣮૷ৄࡉ IPHFQZ features = [ "age", "pclass", "family_size", "fare", "sibsp", "parch", "cabin" ] params_lgb = { 'boosting_type': 'gbdt', 'objective': 'fair', 'metric': 'fair', 'num_boost_round': 20000, 'early_stopping_rounds': 1000, 'verbose': 1000, 'random_state': 999 } cv = { 'method': 'KFold', 'n_splits': 5, 'random_state': 42, 'shuffle': True, } ʙߴ଎ͳ1%$"Λ໨ࢦͯ͠ʙ features = [ "age", "pclass", "family_size", "fare", "sibsp", "parch", "cabin" ] params_lgb = { 'boosting_type': 'gbdt', 'objective': 'fair', 'metric': 'fair', 'num_boost_round': 20000, 'early_stopping_rounds': 1000, 'verbose': 1000, 'random_state': 999 } cv = { 'method': 'GroupKFold', 'n_splits': 5, 'random_state': 42, 'shuffle': True, 'cv_target':'user' } IPHFQZ DWͷهࡌΛมߋ 
    QZUIPO
    IPHFQZ ؆୯
59.  
    ࣮૷ৄࡉ IPHFQZ features = [ "age", "pclass", "family_size", "fare", "sibsp", "parch", "cabin" ] params_lgb = { 'boosting_type': 'gbdt', 'objective': 'fair', 'metric': 'fair', 'num_boost_round': 20000, 'early_stopping_rounds': 1000, 'verbose': 1000, 'random_state': 999 } cv = { 'method': 'KFold', 'n_splits': 5, 'random_state': 42, 'shuffle': True, } ʙߴ଎ͳ1%$"Λ໨ࢦͯ͠ʙ features = [ "age", "pclass", "family_size", "fare", "sibsp", "parch", "cabin" ] params_lgb = { 'boosting_type': 'gbdt', 'objective': 'fair', 'metric': 'fair', 'num_boost_round': 20000, 'early_stopping_rounds': 1000, 'verbose': 1000, 'random_state': 999 } cv = { 'method': 'GroupKFold', 'n_splits': 5, 'random_state': 42, 'shuffle': True, 'cv_target':'user' } IPHFQZ DWͷهࡌΛมߋ 
    QZUIPO
    IPHFQZ GFBUVSF
    JNQPSUBODFʹ͍ͭͯ
60.  
    ࣮૷ৄࡉ ֶशͱಉ࣌ʹGFBUVSF
    JNQPSUBODF͕ը૾ϑΝΠϧͱͯ͠ग़ྗ͞ΕΔ ʙߴ଎ͳ1%$"Λ໨ࢦͯ͠ʙ
61.  
    ࣮૷ৄࡉ ֶशͱಉ࣌ʹGFBUVSF
    JNQPSUBODF͕ը૾ϑΝΠϧͱͯ͠ग़ྗ͞ΕΔ ʙߴ଎ͳ1%$"Λ໨ࢦͯ͠ʙ ͜ͷંΕઢάϥϑ͕ద੾Ͱ͸ͳ͍ͷ͸ঝ஌͓ͯ͠Γ·͢ʜ
    ࠓ͸ΦϨΦϨӡ༻ͳͷͰ͜ͷ··์ஔ͓ͯ͠Γ·ͯ͠
    Ͳ͏͔ࢗͣ͞ʹ͓ئ͍͠·͢
62.  
    ࣮૷ৄࡉ ֶशͱಉ࣌ʹGFBUVSF
    JNQPSUBODF͕ը૾ϑΝΠϧͱͯ͠ग़ྗ͞ΕΔ ʙߴ଎ͳ1%$"Λ໨ࢦͯ͠ʙ ಛ௃ྔ
63.  
    ࣮૷ৄࡉ ֶशͱಉ࣌ʹGFBUVSF
    JNQPSUBODF͕ը૾ϑΝΠϧͱͯ͠ग़ྗ͞ΕΔ ʙߴ଎ͳ1%$"Λ໨ࢦͯ͠ʙ GPMEຖͷฏۉͱඪ४ภࠩ
64.  
    ࣮૷ৄࡉ ֶशͱಉ࣌ʹGFBUVSF
    JNQPSUBODF͕ը૾ϑΝΠϧͱͯ͠ग़ྗ͞ΕΔ ʙߴ଎ͳ1%$"Λ໨ࢦͯ͠ʙ มಈ܎਺ʢඪ४ภࠩฏۉʣ
65.  
    ࣮૷ৄࡉ ֶशͱಉ࣌ʹGFBUVSF
    JNQPSUBODF͕ը૾ϑΝΠϧͱͯ͠ग़ྗ͞ΕΔ ʙߴ଎ͳ1%$"Λ໨ࢦͯ͠ʙ ͜ͷ͋ͨΓͷಛ௃ྔ͸࡟ͬͯྑͦ͞͏
    ͱ͍͏צॴ͕௫ΊΔ
66.  
    ࣮૷ৄࡉ ֶशͱಉ࣌ʹGFBUVSF
    JNQPSUBODF͕ը૾ϑΝΠϧͱͯ͠ग़ྗ͞ΕΔ ʙߴ଎ͳ1%$"Λ໨ࢦͯ͠ʙ ͜ͷ͋ͨΓͷಛ௃ྔ͸࡟ͬͯྑͦ͞͏
    ͱ͍͏צॴ͕௫ΊΔ 1%$"଎͘ճͤͦ͏ײ
67.  ʙҙࣝ͠ͳͯ͘΋؅ཧͰ͖Δ࢓૊Έͷ޻෉ʙ
68.  
    ࣮૷ৄࡉ IPHFQZΛ࣮ߦͨ͋͠ͱͷϑΥϧμɾϑΝΠϧ ʙҙࣝ͠ͳͯ͘΋؅ཧͰ͖Δ࢓૊Έʙ
69.  
    ࣮૷ৄࡉ IPHFQZΛ࣮ߦͨ͋͠ͱͷϑΥϧμɾϑΝΠϧ QMHC@@ ʙҙࣝ͠ͳͯ͘΋؅ཧͰ͖Δ࢓૊Έʙ
70.  
    ࣮૷ৄࡉ IPHFQZΛ࣮ߦͨ͋͠ͱͷϑΥϧμɾϑΝΠϧ QMHC@@ 1VCMJDͷείΞ
    ʢʣ ʙҙࣝ͠ͳͯ͘΋؅ཧͰ͖Δ࢓૊Έʙ
71.  
    ࣮૷ৄࡉ IPHFQZΛ࣮ߦͨ͋͠ͱͷϑΥϧμɾϑΝΠϧ QMHC@@ ֶशΛҰҙʹಛఆ͢Δ໊લ ʙҙࣝ͠ͳͯ͘΋؅ཧͰ͖Δ࢓૊Έʙ
72.  
    ࣮૷ৄࡉ IPHFQZΛ࣮ߦͨ͋͠ͱͷϑΥϧμɾϑΝΠϧ QMHC@@ ΞϧΰϦζϜͷࣝผࢠ ʙҙࣝ͠ͳͯ͘΋؅ཧͰ͖Δ࢓૊Έʙ
73.  
    ࣮૷ৄࡉ IPHFQZΛ࣮ߦͨ͋͠ͱͷϑΥϧμɾϑΝΠϧ QMHC@@ IPHFQZΛ࣮ߦͨ͠೔࣌ ʙҙࣝ͠ͳͯ͘΋؅ཧͰ͖Δ࢓૊Έʙ
74.  
    ࣮૷ৄࡉ IPHFQZΛ࣮ߦͨ͋͠ͱͷϑΥϧμɾϑΝΠϧ ʙҙࣝ͠ͳͯ͘΋؅ཧͰ͖Δ࢓૊Έʙ
75.  
    ࣮૷ৄࡉ IPHFQZΛ࣮ߦͨ͋͠ͱͷϑΥϧμɾϑΝΠϧ ʙҙࣝ͠ͳͯ͘΋؅ཧͰ͖Δ࢓૊Έʙ w ಛ௃ྔ
    w ύϥϝʔλ
    w GFBUVSF
    JNQPSUBODF
    w ֶशϩά
    w Ϟσϧ
    w ਪ࿦ϑΝΠϧ
    ͕อଘ͞ΕΔ
76.  
    ࣮૷ৄࡉ IPHFQZΛ࣮ߦͨ͋͠ͱͷϑΥϧμɾϑΝΠϧ ʙҙࣝ͠ͳͯ͘΋؅ཧͰ͖Δ࢓૊Έʙ w ಛ௃ྔ
    w ύϥϝʔλ
    w GFBUVSF
    JNQPSUBODF
    w ֶशϩά
    w Ϟσϧ
    w ਪ࿦ϑΝΠϧ
    ͕อଘ͞ΕΔ ҙࣝͯ͠ͳ͍͚Ͳউखʹ؅ཧͰ͖ͯΔ
77.  
    ࣮૷ৄࡉ IPHFQZΛ࣮ߦͨ͋͠ͱͷϑΥϧμɾϑΝΠϧ ʙҙࣝ͠ͳͯ͘΋؅ཧͰ͖Δ࢓૊Έʙ w ಛ௃ྔ
    w ύϥϝʔλ
    w GFBUVSF
    JNQPSUBODF
    w ֶशϩά
    w Ϟσϧ
    w ਪ࿦ϑΝΠϧ
    ͕อଘ͞ΕΔ
78. ·ͱΊ 
79.  ·ͱΊ w ֶशɾਪ࿦ύΠϓϥΠϯ͍͍ͧʂ
     ύΠϓϥΠϯΛߏங͢Δ͜ͱͰԼهͷΑ͏ͳϝϦοτ͕͋Γʢࠓͷॴʣݸਓత ʹ͸ΊͬͪΌྑ͍ɻ
    ‣ ࠶ݱੑ
    ‣ ߴ଎ͳ1%$"
    ‣ ॾʑͷ؅ཧ
    w ্هͷΑ͏ͳ͜ͱ͕୲อ͞ΕΔͷͰ৺ཧత҆શੑ΋
80.  ϫΠ͸͜͏΍ͬͯΔͥʂ
    ͱ͍͏ͷ͕͋Ε͹ੋඇ࠙਌ձͰڭ͑ͯԼ͍͞ʂ
81.  ͝ਗ਼ௌ͋Γ͕ͱ͏͍͟͝·ͨ͠