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
ֶशɾਪύΠϓϥΠϯΛߏங͢Δ্Ͱେʹ͍ͯ͠Δ͜ͱ$POOFIJUP*ODᖒরੳίϯϖ-5ձ
View Slide
͜Μʹͪʂ
ࠓൃද͢Δ͜ͱԼهͱएׯॏෳ͠·͢IUUQTTQFBLFSEFDLDPNUBLBQZEFUBGFOYJLPOQFOJPJUFUF[IFOHMJBOHHVBOMJOJQJCJTJUFJSVRVBOSFOMFJOJDIVBOFUBJYJBOHJ
ΞδΣϯμ ࣗݾհ ύΠϓϥΠϯ͕ͳ͔ͬͨ࣌ͷπϥϛ େʹ͍ͯ͠Δ͜ͱ‣ ֶशͷ࠶ݱੑ‣ 1%$"ͷߴԽ‣ ҙࣝ͠ͳͯ͘ཧͰ͖ΔΈ ࣮ৄࡉ
ࣗݾհ
ࣗݾհ໊લɿᖒরʢ/P[BXB5BLBOPCVʣॴଐɿίωώτגࣜձࣾɹɹɿ͔ͨͺ͍!UBLBQZw ʙίωώτʹ.-ΤϯδχΞͱͯ͠+0*/w ػցֶशʢ/-1ɺਪનγεςϜʣΛϝΠϯʹΓͭͭΠϯϑϥʢ"84ʣษڧதw ,BHHMFͨ͠ΓɺϒϩάʢIUUQTXXXUBLBQZXPSLʣॻ͍ͨΓɺٿͨ͠Γɺϥʔϝϯ৯ͨΓ͍ͯ͠·͢w ΦεεϝͷυϥϜࣜચ୕ػ͋ͬͨΒڭ͍͑ͯͩ͘͞
ࣗݾհ˞ʮӾཡʯʮར༻ऀʯϝσΟΞͱΞϓϦͷ߹ܭʢ݄݄ͷฏۉʣ˞ʮϚϚ͚/PΞϓϦʯ݄Πϯςʔδௐɹௐࠪରɿ৷தʙ̎ࡀ̌ϲ݄ͷࢠڙΛ࣋ͭঁੑ O Λநग़˞*OTUBHSBNͷϑΥϩϫʔɺ'BDFCPPLͷ͍͍Ͷɺ-*/&ͷͱͩͪͷ߹ܭ ݄࣌ϚϚϦ ΞϓϦɾ8FC 4/4 *OTUBHSBNɾ-*/&ɾ'BDFCPPLهࣄϚϚಉ࢜ͰΈΛ૬ஊ͠߹͏2"ίϛϡχςΟΛத৺ʹϢʔβʔΛ֦େ͍ͯ͠·͢ʮϚϚϦʯͰϢʔβʔಉ͕࢜ͲΜͲΜܨ͕͍ͬͯ·͢ϚϚͷੜ׆ʹཱͭهࣄΛ෯͍δϟϯϧͰ৴͍ͯ͠·͢ϚϚ͚/P̍ΞϓϦʹબग़ ਓͷϚϚ͕બͿʮݱࡏ͍ͬͯΔΞϓϦʯʹͯɺ߲ ଞͷϚϚʹΦεεϝ͍ͨ͠ɺೝɺར༻ɺརศੑɺײͰ̍ҐΛ֫ಘ͠·ͨ͠هࣄ6,000 هࣄҎ্ྦྷܭϑΝϯ85 ສਓ˞݄ؒӾཡ1.5ԯճ˞݄ؒར༻ऀ650ສਓ˞˞lΈzͱzڞײzΛ࣠ʹϚϚʹدΓఴ͍ΞϓϦɾ8FCɾ4/4ͱଟ֯తʹαʔϏεΛల։͍ͯ͠·͢
ࣗݾհ0450,000900,0001,350,0001,800,0002014/42014/52014/62014/72014/82014/92014/102014/112014/122015/12015/22015/32015/42015/52015/62015/72015/82015/92015/102015/112015/122016/12016/22016/32016/42016/52016/62016/72016/82016/92016/102016/112016/122017/12017/22017/32017/42017/52017/62017/72017/82017/92017/102017/112017/122018/12018/22018/32018/42018/52018/62018/72018/8݄ؒߘ150ສ݅िʹҎ্ىಈ͢ΔΞΫςΟϒϢʔβʔ50ਓʹਓ57$.์өΞϓϦ૯%-ສਓʹਓਓʹਓਓʹਓਓʹਓ˞˞ʮϚϚϦʯͷग़࢈༧ఆΛઃఆͨ͠Ϣʔβʔͱɺްੜ࿑ಇলൃදʮਓޱಈଶ౷ܭʯͷग़ੜ͔Βࢉग़˞िʹճҎ্ىಈ͢ΔϢʔβʔ˞ʹग़࢈ͨ͠ϚϚͷʮਓʹਓʯ͕ϚϚϦΛར༻தຊ࠷େڃنΛތΔϒϥϯυͱ͍ͯ͠·͢˞
ύΠϓϥΠϯ͕ͳ͔ͬͨ࣌ͷπϥϛ
ύΠϓϥΠϯ͕ͳ͔ͬͨ࣌ͷπϥϛΊͬͪΌྑ͍είΞͷOPUFCPPL͕ˣ͜ͷOPUFCPPLΛ%VQMJDBUFͯ͠ɺͬͱྑ͍Ϟσϧ࡞ͬͪΌ͏ͧʂ
ҰํɺOPUFCPPLͷதʜ
ύΠϓϥΠϯ͕ͳ͔ͬͨ࣌ͷπϥϛ<> import numpy as npimport pandas as pdOPUFCPPLͷதɾɾɾ<> hogehoge<> hogehoge<> hogehogeɾɾɾ
ύΠϓϥΠϯ͕ͳ͔ͬͨ࣌ͷπϥϛ<> import numpy as npimport pandas as pdOPUFCPPLͷதɾɾɾ<> hogehoge<> hogehoge<> hogehogeɾɾɾηϧͷ࣮ߦॱ͕ͪ͝Όͪ͝Όˣ࠶ݱੑ͕ͳ͍
ύΠϓϥΠϯ͕ͳ͔ͬͨ࣌ͷπϥϛ<> import numpy as npimport pandas as pdOPUFCPPLͷதɾɾɾ<> submission.to_csv('submission.csv', index=False)
ύΠϓϥΠϯ͕ͳ͔ͬͨ࣌ͷπϥϛ<> import numpy as npimport pandas as pdOPUFCPPLͷதɾɾɾ<> submission.to_csv('submission.csv', index=False)ηϧ͕ଟ͘ɺಉ͡ܭࢉΛෳ࣮ߦ͠ͳ͚ΕͳΒͳ͍ˣ1%$"͕͘ͳΓ͕ͪ
ύΠϓϥΠϯ͕ͳ͔ͬͨ࣌ͷπϥϛdOPUFCPPL-JHIU#(.@TDPSF@$PQZJQZOCdOPUFCPPL-JHIU#(.@TDPSF@$PQZJQZOCdOPUFCPPL-JHIU#(.@TDPSF@$PQZJQZOCdOPUFCPPL-JHIU#(.@TDPSF@$PQZJQZOCdOPUFCPPL-JHIU#(.@TDPSF@$PQZJQZOCdOPUFCPPL-JHIU#(.@TDPSF@$PQZJQZOCdOPUFCPPL-JHIU#(.@TDPSF@$PQZJQZOCdOPUFCPPL-JHIU#(.@TDPSF@$PQZJQZOCdOPUFCPPL-JHIU#(.@TDPSF@JQZOCdOPUFCPPL-JHIU#(.@TDPSF@JQZOCdOPUFCPPL-JHIU#(.@TDPSF@JQZOCdOPUFCPPL-JHIU#(.@TDPSF@JQZOCdOPUFCPPL-JHIU#(.@TDPSF@JQZOCdOPUFCPPL-JHIU#(.@TDPSF@JQZOCdOPUFCPPL-JHIU#(.@TDPSF@JQZOCdOPUFCPPL-JHIU#(.@TDPSF@JQZOCʜʜʜʜʜ
ύΠϓϥΠϯ͕ͳ͔ͬͨ࣌ͷπϥϛdOPUFCPPL-JHIU#(.@TDPSF@$PQZJQZOCdOPUFCPPL-JHIU#(.@TDPSF@$PQZJQZOCdOPUFCPPL-JHIU#(.@TDPSF@$PQZJQZOCdOPUFCPPL-JHIU#(.@TDPSF@$PQZJQZOCdOPUFCPPL-JHIU#(.@TDPSF@$PQZJQZOCdOPUFCPPL-JHIU#(.@TDPSF@$PQZJQZOCdOPUFCPPL-JHIU#(.@TDPSF@$PQZJQZOCdOPUFCPPL-JHIU#(.@TDPSF@$PQZJQZOCdOPUFCPPL-JHIU#(.@TDPSF@JQZOCdOPUFCPPL-JHIU#(.@TDPSF@JQZOCdOPUFCPPL-JHIU#(.@TDPSF@JQZOCdOPUFCPPL-JHIU#(.@TDPSF@JQZOCdOPUFCPPL-JHIU#(.@TDPSF@JQZOCdOPUFCPPL-JHIU#(.@TDPSF@JQZOCdOPUFCPPL-JHIU#(.@TDPSF@JQZOCdOPUFCPPL-JHIU#(.@TDPSF@JQZOCʜʜʜʜʜOPUFCPPL͕ࡍݶͳ͘૿৩͠ཧ͕ΊͪΌͪ͘Όࡶʹ
͜ΕΒͷπϥϛΛղফ͘͢ɾOPUFCPPLʹνϟϨϯδˣͦΜͳதͰେʹ͍ͯ͠Δ͜ͱΛ͓͑͠·͢˞&%"ͳͲOPUFCPPLΛ༻
͜ͷຊͰ͢
େʹ͍ͯ͠Δ͜ͱ
ͦͷʙֶशͷ࠶ݱੑʙ
ͦͷʙֶशͷ࠶ݱੑʙϙΠϯτ̏ͭ
େʹ͍ͯ͠Δ͜ͱw ͲΜͳಛྔΛͬͯʙֶशͷ࠶ݱੑʙ
େʹ͍ͯ͠Δ͜ͱw ͲΜͳಛྔΛͬͯw ͲΜͳύϥϝʔλʔΛͬͯʙֶशͷ࠶ݱੑʙ
େʹ͍ͯ͠Δ͜ͱw ͲΜͳಛྔΛͬͯw ͲΜͳύϥϝʔλʔΛͬͯw ͲΜͳ$7Λֶͬͯश͔ͨ͠Λอଘ͓ͯ͘͠ɻʙֶशͷ࠶ݱੑʙ
େʹ͍ͯ͠Δ͜ͱw ͲΜͳಛྔΛͬͯw ͲΜͳύϥϝʔλʔΛͬͯw ͲΜͳ$7Λֶͬͯश͔ͨ͠Λอଘ͓ͯ͘͠ɻʙֶशͷ࠶ݱੑʙ%0//"อଘଇ
ͦͷʙߴͳ1%$"Λࢦͯ͠ʙ
େʹ͍ͯ͠Δ͜ͱw ͭͷQZUIPOεΫϦϓτͰ%0//"ΛঠѲ ༻͢Δಛྔ ༻͢Δύϥϝʔλ ༻͢Δ$7ʙߴͳ1%$"Λࢦͯ͠ʙ
େʹ͍ͯ͠Δ͜ͱw ͭͷQZUIPOεΫϦϓτͰ%0//"ΛঠѲ ༻͢Δಛྔ ༻͢Δύϥϝʔλ ༻͢Δ$7w GFBUVSFJNQPSUBODFͷݟ͑ΔԽ ࣍ͷ࣮ݧͷצॴΛ͔ͭΊΔΑ͏ʹʙߴͳ1%$"Λࢦͯ͠ʙ
ͦͷʙҙࣝ͠ͳͯ͘ཧͰ͖ΔΈʙ
େʹ͍ͯ͠Δ͜ͱ ʙҙࣝ͠ͳͯ͘ཧͰ͖ΔΈʙw %0///"ϞσϧϩάϑΝΠϧΛɺ࣮ߦ͝ͱʹʮ࣌ʯTV⒏YΛ͚ͭͨσΟϨΫτϦɾϑΝΠϧͰࣗಈཧ ༻͢Δಛྔ ύϥϝʔλ $7 ֶशͨ͠Ϟσϧ ֶशϩάϑΝΠϧ GFBUVSFJNQPSUBODF
େʹ͍ͯ͠Δ͜ͱ ʙҙࣝ͠ͳͯ͘ཧͰ͖ΔΈʙw %0///"ϞσϧϩάϑΝΠϧΛɺ࣮ߦ͝ͱʹʮ࣌ʯTV⒏YΛ͚ͭͨσΟϨΫτϦɾϑΝΠϧͰࣗಈཧ ༻͢Δಛྔ ύϥϝʔλ $7 ֶशͨ͠Ϟσϧ ֶशϩάϑΝΠϧ GFBUVSFJNQPSUBODFw 1VCMJDείΞΛσΟϨΫτϦͷ1SFpYʹ͚Δ͜ͱͰɺ1VCMJDͱMPDBMͷείΞͷରԠΛ͔Γ͘͢ʢৄࡉޙ΄Ͳʣ
େʹ͍ͯ͠Δ͜ͱ ʙҙࣝ͠ͳͯ͘ཧͰ͖ΔΈʙw %0///"ϞσϧϩάϑΝΠϧΛɺ࣮ߦ͝ͱʹʮ࣌ʯTV⒏YΛ͚ͭͨσΟϨΫτϦɾϑΝΠϧͰࣗಈཧ ༻͢Δಛྔ ύϥϝʔλ $7 ֶशͨ͠Ϟσϧ ֶशϩάϑΝΠϧ GFBUVSFJNQPSUBODFw 1VCMJDείΞΛσΟϨΫτϦͷ1SFpYʹ͚Δ͜ͱͰɺ1VCMJDͱMPDBMͷείΞͷରԠΛ͔Γ͘͢ʢৄࡉޙ΄ͲʣৄࡉΛݟ͍͖ͯ·͢
࣮ৄࡉ
ྫʣIPHFQZfeatures = ["age","pclass","family_size","fare","sibsp","parch","cabin"]params_lgb = {'boosting_type': 'gbdt','objective': 'fair','metric': 'fair','num_boost_round': 20000,'early_stopping_rounds': 1000,'verbose': 1000,'random_state': 999}cv = {'method': 'KFold','n_splits': 5,'random_state': 42,'shuffle': True,}runner = Runner(run_name, ModelLGB, features, dataset.get('target'), params_lgb, cv, FEATURE_DIR_NAME, MODEL_DIR_NAME)runner.run_train_cv() # ֶशrunner.run_predict_cv() # ਪSubmission.create_submission(run_name) # submit࡞
ྫʣIPHFQZfeatures = ["age","pclass","family_size","fare","sibsp","parch","cabin"]params_lgb = {'boosting_type': 'gbdt','objective': 'fair','metric': 'fair','num_boost_round': 20000,'early_stopping_rounds': 1000,'verbose': 1000,'random_state': 999}cv = {'method': 'KFold','n_splits': 5,'random_state': 42,'shuffle': True,}runner = Runner(run_name, ModelLGB, features, dataset.get('target'), params_lgb, cv, FEATURE_DIR_NAME, MODEL_DIR_NAME)runner.run_train_cv() # ֶशrunner.run_predict_cv() # ਪSubmission.create_submission(run_name) # submit࡞جຊతʹ1%$"Λճ͢ͱ͖ʹ͍͡Δͷ͜ͷIPHFQZͷΈʹ͢Δ
ྫʣIPHFQZfeatures = ["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() # ֶशrunner.run_predict_cv() # ਪSubmission.create_submission(run_name) # submit࡞ֶशʹ༻͢Δಛྔ
ྫʣIPHFQZfeatures = ["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() # ֶशrunner.run_predict_cv() # ਪSubmission.create_submission(run_name) # submit࡞ϋΠύʔύϥϝʔλ
ྫʣIPHFQZfeatures = ["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() # ֶशrunner.run_predict_cv() # ਪSubmission.create_submission(run_name) # submit࡞$7ͷઃఆ
ྫʣIPHFQZfeatures = ["age","pclass","family_size","fare","sibsp","parch","cabin"]params_lgb = {'boosting_type': 'gbdt','objective': 'fair','metric': 'fair','num_boost_round': 20000,'early_stopping_rounds': 1000,'verbose': 1000,'random_state': 999}cv = {'method': 'KFold','n_splits': 5,'random_state': 42,'shuffle': True,}runner = Runner(run_name, ModelLGB, features, dataset.get('target'), params_lgb, cv, FEATURE_DIR_NAME, MODEL_DIR_NAME)runner.run_train_cv() # ֶशrunner.run_predict_cv() # ਪSubmission.create_submission(run_name) # submit࡞ֶशɾਪɾTVC࡞
ྫʣIPHFQZfeatures = ["age","pclass","family_size","fare","sibsp","parch","cabin"]params_lgb = {'boosting_type': 'gbdt','objective': 'fair','metric': 'fair','num_boost_round': 20000,'early_stopping_rounds': 1000,'verbose': 1000,'random_state': 999}cv = {'method': 'KFold','n_splits': 5,'random_state': 42,'shuffle': True,}runner = Runner(run_name, ModelLGB, features, dataset.get('target'), params_lgb, cv, FEATURE_DIR_NAME, MODEL_DIR_NAME)runner.run_train_cv() # ֶशrunner.run_predict_cv() # ਪSubmission.create_submission(run_name) # submit࡞ֶशɾਪɾTVC࡞w ࠶ݱੑΛ୲อ͢Δw ߴͳ1$%"Λճͨ͢Ίͷw ҙࣝ͠ͳͯ͘ॾʑ͕ཧͰ͖Δʹ͍͓ͭͯ͠͠·͢
ʙֶशͷ࠶ݱੑΛ୲อ͢ΔͨΊͷʙ
࣮ৄࡉIPHFQZʙֶशͷ࠶ݱੑʙfeatures = ["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,}
࣮ৄࡉIPHFQZʙֶशͷ࠶ݱੑʙfeatures = ["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,}QZUIPOIPHFQZ
࣮ৄࡉIPHFQZʙֶशͷ࠶ݱੑʙfeatures = ["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@QBSBNKTPOQZUIPOIPHFQZ
࣮ৄࡉIPHFQZʙֶशͷ࠶ݱੑʙfeatures = ["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@QBSBNKTPOQZUIPOIPHFQZIPHFQZΛ࣮ߦ͢Δ͜ͱʹΑΓࣗಈతʹKTPOϑΝΠϧ͕ੜ͞Ε༻ͨ͠ಛྔɾύϥϝʔλʔͳͲ͕શͯอଘ͞ΕΔ
࣮ৄࡉIPHFQZʙֶशͷ࠶ݱੑʙfeatures = ["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@QBSBNKTPOQZUIPOIPHFQZ͜Εͧ࠶ݱੑ
ʙߴͳ1%$"Λճͨ͢Ίͷʙ
࣮ৄࡉIPHFQZfeatures = ["age","pclass","family_size","fare","sibsp","parch","cabin"]params_lgb = {'boosting_type': 'gbdt','objective': 'fair','metric': 'fair','num_boost_round': 20000,'early_stopping_rounds': 1000,'verbose': 1000,'random_state': 999}cv = {'method': 'KFold','n_splits': 5,'random_state': 42,'shuffle': True,}ʙߴͳ1%$"Λࢦͯ͠ʙ
࣮ৄࡉIPHFQZfeatures = ["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,}ʙߴͳ1%$"Λࢦͯ͠ʙಛྔΛݮΒֶͯ͠श͍ͤͨ͞
࣮ৄࡉIPHFQZfeatures = ["age","pclass","family_size","fare","sibsp","parch","cabin"]params_lgb = {'boosting_type': 'gbdt','objective': 'fair','metric': 'fair','num_boost_round': 20000,'early_stopping_rounds': 1000,'verbose': 1000,'random_state': 999}cv = {'method': 'KFold','n_splits': 5,'random_state': 42,'shuffle': True,}ʙߴͳ1%$"Λࢦͯ͠ʙfeatures = ["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 ରͷಛྔΛίϝϯτΞτ
࣮ৄࡉIPHFQZfeatures = ["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,}ʙߴͳ1%$"Λࢦͯ͠ʙfeatures = ["age","pclass",# "family_size","fare","sibsp","parch","cabin"]params_lgb = {'boosting_type': 'gbdt','objective': 'fair','metric': 'fair','num_boost_round': 20000,'early_stopping_rounds': 1000,'verbose': 1000,'random_state': 999}cv = {'method': 'KFold','n_splits': 5,'random_state': 42,'shuffle': True,}IPHFQZQZUIPOIPHFQZରͷಛྔΛίϝϯτΞτ
࣮ৄࡉIPHFQZfeatures = ["age","pclass","family_size","fare","sibsp","parch","cabin"]params_lgb = {'boosting_type': 'gbdt','objective': 'fair','metric': 'fair','num_boost_round': 20000,'early_stopping_rounds': 1000,'verbose': 1000,'random_state': 999}cv = {'method': 'KFold','n_splits': 5,'random_state': 42,'shuffle': True,}ʙߴͳ1%$"Λࢦͯ͠ʙfeatures = ["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,}IPHFQZQZUIPOIPHFQZରͷಛྔΛίϝϯτΞτ؆୯
࣮ৄࡉIPHFQZfeatures = ["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,}ʙߴͳ1%$"Λࢦͯ͠ʙ$7มֶ͑ͯश͍ͤͨ͞
࣮ৄࡉIPHFQZfeatures = ["age","pclass","family_size","fare","sibsp","parch","cabin"]params_lgb = {'boosting_type': 'gbdt','objective': 'fair','metric': 'fair','num_boost_round': 20000,'early_stopping_rounds': 1000,'verbose': 1000,'random_state': 999}cv = {'method': 'KFold','n_splits': 5,'random_state': 42,'shuffle': True,}ʙߴͳ1%$"Λࢦͯ͠ʙfeatures = ["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
࣮ৄࡉIPHFQZfeatures = ["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,}ʙߴͳ1%$"Λࢦͯ͠ʙfeatures = ["age","pclass","family_size","fare","sibsp","parch","cabin"]params_lgb = {'boosting_type': 'gbdt','objective': 'fair','metric': 'fair','num_boost_round': 20000,'early_stopping_rounds': 1000,'verbose': 1000,'random_state': 999}cv = {'method': 'GroupKFold','n_splits': 5,'random_state': 42,'shuffle': True,'cv_target':'user'}IPHFQZ DWͷهࡌΛมߋQZUIPOIPHFQZ
࣮ৄࡉIPHFQZfeatures = ["age","pclass","family_size","fare","sibsp","parch","cabin"]params_lgb = {'boosting_type': 'gbdt','objective': 'fair','metric': 'fair','num_boost_round': 20000,'early_stopping_rounds': 1000,'verbose': 1000,'random_state': 999}cv = {'method': 'KFold','n_splits': 5,'random_state': 42,'shuffle': True,}ʙߴͳ1%$"Λࢦͯ͠ʙfeatures = ["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 DWͷهࡌΛมߋQZUIPOIPHFQZ؆୯
࣮ৄࡉIPHFQZfeatures = ["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,}ʙߴͳ1%$"Λࢦͯ͠ʙfeatures = ["age","pclass","family_size","fare","sibsp","parch","cabin"]params_lgb = {'boosting_type': 'gbdt','objective': 'fair','metric': 'fair','num_boost_round': 20000,'early_stopping_rounds': 1000,'verbose': 1000,'random_state': 999}cv = {'method': 'GroupKFold','n_splits': 5,'random_state': 42,'shuffle': True,'cv_target':'user'}IPHFQZ DWͷهࡌΛมߋQZUIPOIPHFQZGFBUVSFJNQPSUBODFʹ͍ͭͯ
࣮ৄࡉֶशͱಉ࣌ʹGFBUVSFJNQPSUBODF͕ը૾ϑΝΠϧͱͯ͠ग़ྗ͞ΕΔʙߴͳ1%$"Λࢦͯ͠ʙ
࣮ৄࡉֶशͱಉ࣌ʹGFBUVSFJNQPSUBODF͕ը૾ϑΝΠϧͱͯ͠ग़ྗ͞ΕΔʙߴͳ1%$"Λࢦͯ͠ʙ͜ͷંΕઢάϥϑ͕దͰͳ͍ͷঝ͓ͯ͠Γ·͢ʜࠓΦϨΦϨӡ༻ͳͷͰ͜ͷ··์ஔ͓ͯ͠Γ·ͯ͠Ͳ͏͔ͣ͞ʹ͓ئ͍͠·͢
࣮ৄࡉֶशͱಉ࣌ʹGFBUVSFJNQPSUBODF͕ը૾ϑΝΠϧͱͯ͠ग़ྗ͞ΕΔʙߴͳ1%$"Λࢦͯ͠ʙಛྔ
࣮ৄࡉֶशͱಉ࣌ʹGFBUVSFJNQPSUBODF͕ը૾ϑΝΠϧͱͯ͠ग़ྗ͞ΕΔʙߴͳ1%$"Λࢦͯ͠ʙGPMEຖͷฏۉͱඪ४ภࠩ
࣮ৄࡉֶशͱಉ࣌ʹGFBUVSFJNQPSUBODF͕ը૾ϑΝΠϧͱͯ͠ग़ྗ͞ΕΔʙߴͳ1%$"Λࢦͯ͠ʙมಈʢඪ४ภࠩฏۉʣ
࣮ৄࡉֶशͱಉ࣌ʹGFBUVSFJNQPSUBODF͕ը૾ϑΝΠϧͱͯ͠ग़ྗ͞ΕΔʙߴͳ1%$"Λࢦͯ͠ʙ͜ͷ͋ͨΓͷಛྔͬͯྑͦ͞͏ͱ͍͏צॴ͕௫ΊΔ
࣮ৄࡉֶशͱಉ࣌ʹGFBUVSFJNQPSUBODF͕ը૾ϑΝΠϧͱͯ͠ग़ྗ͞ΕΔʙߴͳ1%$"Λࢦͯ͠ʙ͜ͷ͋ͨΓͷಛྔͬͯྑͦ͞͏ͱ͍͏צॴ͕௫ΊΔ1%$"͘ճͤͦ͏ײ
ʙҙࣝ͠ͳͯ͘ཧͰ͖ΔΈͷʙ
࣮ৄࡉIPHFQZΛ࣮ߦͨ͋͠ͱͷϑΥϧμɾϑΝΠϧʙҙࣝ͠ͳͯ͘ཧͰ͖ΔΈʙ
࣮ৄࡉIPHFQZΛ࣮ߦͨ͋͠ͱͷϑΥϧμɾϑΝΠϧQMHC@@ʙҙࣝ͠ͳͯ͘ཧͰ͖ΔΈʙ
࣮ৄࡉIPHFQZΛ࣮ߦͨ͋͠ͱͷϑΥϧμɾϑΝΠϧQMHC@@1VCMJDͷείΞʢʣʙҙࣝ͠ͳͯ͘ཧͰ͖ΔΈʙ
࣮ৄࡉIPHFQZΛ࣮ߦͨ͋͠ͱͷϑΥϧμɾϑΝΠϧQMHC@@ֶशΛҰҙʹಛఆ͢Δ໊લʙҙࣝ͠ͳͯ͘ཧͰ͖ΔΈʙ
࣮ৄࡉIPHFQZΛ࣮ߦͨ͋͠ͱͷϑΥϧμɾϑΝΠϧQMHC@@ΞϧΰϦζϜͷࣝผࢠʙҙࣝ͠ͳͯ͘ཧͰ͖ΔΈʙ
࣮ৄࡉIPHFQZΛ࣮ߦͨ͋͠ͱͷϑΥϧμɾϑΝΠϧQMHC@@IPHFQZΛ࣮ߦͨ࣌͠ʙҙࣝ͠ͳͯ͘ཧͰ͖ΔΈʙ
࣮ৄࡉIPHFQZΛ࣮ߦͨ͋͠ͱͷϑΥϧμɾϑΝΠϧʙҙࣝ͠ͳͯ͘ཧͰ͖ΔΈʙw ಛྔw ύϥϝʔλw GFBUVSFJNQPSUBODFw ֶशϩάw Ϟσϧw ਪϑΝΠϧ͕อଘ͞ΕΔ
࣮ৄࡉIPHFQZΛ࣮ߦͨ͋͠ͱͷϑΥϧμɾϑΝΠϧʙҙࣝ͠ͳͯ͘ཧͰ͖ΔΈʙw ಛྔw ύϥϝʔλw GFBUVSFJNQPSUBODFw ֶशϩάw Ϟσϧw ਪϑΝΠϧ͕อଘ͞ΕΔҙࣝͯ͠ͳ͍͚ͲউखʹཧͰ͖ͯΔ
·ͱΊ
·ͱΊw ֶशɾਪύΠϓϥΠϯ͍͍ͧʂ ύΠϓϥΠϯΛߏங͢Δ͜ͱͰԼهͷΑ͏ͳϝϦοτ͕͋ΓʢࠓͷॴʣݸਓతʹΊͬͪΌྑ͍ɻ‣ ࠶ݱੑ‣ ߴͳ1%$"‣ ॾʑͷཧw ্هͷΑ͏ͳ͜ͱ͕୲อ͞ΕΔͷͰ৺ཧత҆શੑ
ϫΠ͜͏ͬͯΔͥʂͱ͍͏ͷ͕͋Εੋඇ࠙ձͰڭ͑ͯԼ͍͞ʂ
͝ਗ਼ௌ͋Γ͕ͱ͏͍͟͝·ͨ͠
takapy
codenote
oldbigbuddha
yoshidashingo
sat
coconala_engineer
nwiizo
sansantech
mtx2s
upura
shibuiwilliam
onecareer_tech
bluesmoon
jnunemaker
garrettdimon
paigeccino
iamctodd
trallard
quramy
brianwarren
mthomps
geoffreycrofte
speakerdeck
gr2m