BigQuery ML を使ってみた話

Slide 1

Slide 1 text

Slide 2

Slide 2 text

©2018 Wantedly, Inc. ✓ Yuya Matsumura (দଜ ༏໵) ✓ Software Engineer (ݕࡧɾਪનΤϯδχΞʁ) ✓ Wantedly, Inc. Recommendation Team (since April 2018) ✓ Interested in Information Retrieval, Machine Learning Self Introduction @yu-ya4 @yu__ya4

Slide 3

Slide 3 text

Slide 4

Slide 4 text

Slide 5

Slide 5 text

Slide 6

Slide 6 text

Slide 7

Slide 7 text

Slide 8

Slide 8 text

Slide 9

Slide 9 text

Slide 10

Slide 10 text

Slide 11

Slide 11 text

Slide 12

Slide 12 text

©2018 Wantedly, Inc. ✓ Google Cloud Next 2018 Ͱൃද ✓ BigQuery + ML (ػցֶश) • طଘͷ SQL πʔϧͱεΩϧͰػցֶश͕Ͱ͖Δ • ػցֶशͷຽओԽ (σʔλΞφϦετɼϏδωενʔϜ΋࢖͑Δʂ) ✓ BigQuery͚ͩͰ׬݁͢ΔͨΊखܰʂ • σʔλΛϩʔΧϧ؀ڥ౳ʹҠಈͤ͞Δඞཁͳ͠ • ։ൃεϐʔυͷ޲্ What is BigQuery ML? https://cloud.google.com/bigquery/docs/bigqueryml-intro

Slide 13

Slide 13 text

©2018 Wantedly, Inc. ✓ ݱࡏαϙʔτ͍ͯ͠ΔϞσϧ͸ҎԼͷ 3 छ • ઢܗճؼϞσϧ (σʔλ͔Β਺஋ͷਪఆ) • ೋ߲ϩδεςΟοΫճؼϞσϧ (σʔλ͔Β true/false Λ൑ఆ) • ଟ߲ϩδεςΟοΫճؼϞσϧ(σʔλΛ3ͭҎ্ͷΫϥεʹ෼ྨ) What is BigQuery ML? https://cloud.google.com/bigquery/docs/bigqueryml-intro

Slide 14

Slide 14 text

Slide 15

Slide 15 text

Slide 16

Slide 16 text

Slide 17

Slide 17 text

Slide 18

Slide 18 text

Slide 19

Slide 19 text

Slide 20

Slide 20 text

Slide 21

Slide 21 text

Slide 22

Slide 22 text

Slide 23

Slide 23 text

Slide 24

Slide 24 text

Slide 25

Slide 25 text

©2018 Wantedly, Inc. Define Job BigQuery ͷ࣮ߦ݁ՌΛ೔࣌Ͱςʔϒϧʹॻ͖ग़͢ export do table :daily_company_scores columns %w[:day, :company_id, :score] mode :update, [:day] end schedule do frequency :daily end run :bq, <

Slide 26

Slide 26 text

Slide 27

Slide 27 text

Slide 28

Slide 28 text

Slide 29

Slide 29 text

©2018 Wantedly, Inc. Overview of architecture BigQuery merge Scheduler ʹొ࿥ Job(ΫΤϦ)ͷ࣮ߦ ΫΤϦͷ࣮ߦ݁ՌΛ BigQuery ΍ RDBʹॻ͖ग़͢ Job ͷهड़(΄΅ SQL) Logging Read Job ϑΝΠϧΛ௥Ճ͠ɼpull-request ʹͯ͠Ϛʔδ ࣗಈͰ Kubernetes Λ׆༻ͨ͠Scheduler ʹొ࿥͞Εఆظ࣮ߦ͞ΕΔ

Slide 30

Slide 30 text

Slide 31

Slide 31 text

Slide 32

Slide 32 text

Slide 33

Slide 33 text

Slide 34

Slide 34 text

©2018 Wantedly, Inc. ✓ ݱࡏͷWantedly Visit ͷ Recommendation Team ࣄ৘ γϯϓϧͳΞΠσΞͰվળͰ͖ͦ͏ͳՕॴ͕ෳ਺࢒͍ͬͯΔ νʔϜ͕খن໛ ࢲΛؚΊ໊ͯ αʔϏε͸খن໛͡Όͳ͍ ͨ͘͞ΜͷϢʔβ ͭͭͷվળʹΊ͍͍ͬͺ͍࣌ؒΛ͔͚͍͍ͯΘ͚Ͱ͸ͳ͍ վળαΠΫϧΛ͢͹΍͘ճ͍ͨ͠ BigQuery in Recommendation Team

Slide 35

Slide 35 text

Slide 36

Slide 36 text

Slide 37

Slide 37 text

Slide 38

Slide 38 text

©2018 Wantedly, Inc. It is often necessary to make choices without sufficient personal experience of the alternatives. In everyday life, we rely on recommendations from other people either by word of mouth, recommendation letters, movie and book reviews printed in newspapers, or general surveys such as Zagat’s restaurant guides. Recommender systems assist and augment this natural social process. ਪનγεςϜͱ͸ [Resnick+ 97] ࣗ෼ͷܦݧ͔ΒͷΈͰ͸͋·ΓΑ͘Θ͔Βͳ͍΋ͷͷத͔ΒɼͲ͏ ͯ͠΋ͲΕ͔Λબ͹ͳ͚Ε͹ͳΒͳ͍ͱ͍͏͜ͱ͸Α͋͘Δɽ͜ͷ Α͏ͳࡍΘͨͨͪ͠͸ɼޱίϛɼਪનঢ়ɼ ৽ฉͷॻධ΍өըධɼϨ ετϥϯΨΠυͳͲͷଞਓ͔ΒͷਪનʹཔΔ͜ͱΛ ೔ৗతʹߦͬͯ ͍Δɽ ৘ใਪનγεςϜͱ͸ɼ͜͏ͨࣾ͠ձͰී௨ʹߦΘΕ͍ͯΔ Ұ࿈ͷߦҝΛิॿͨ͠Γɼଅਐͨ͠Γ͢Δ΋ͷͰ͋Δɽ [Resnick 97] P. Resnick and H. R. Varian. Recommender systems. Communications of the ACM, Vol. 40, No. 3, pp. 56–58, 1997.

Slide 39

Slide 39 text

©2018 Wantedly, Inc. Recommenders: Tools to help identify worthwhile stuff ਪનγεςϜͱ͸ [Konstan+ 03] ਪનγεςϜ: Ձ஋ͷ͋Δ΋ͷ͕ͲΕ͔ಛఆ͢Δͷʹ໾ཱͭಓ۩ [Konstan 03] J. A. Konstan and J. Riedl. Recommender systems: Collaborating in commerce and communities. In Proc. of the SIGCHI Conf. on Human Factors in Computing Systems, Tutorial, 2003.

Slide 40

Slide 40 text

Slide 41

Slide 41 text

Slide 42

Slide 42 text

Slide 43

Slide 43 text

Slide 44

Slide 44 text

Slide 45

Slide 45 text

Slide 46

Slide 46 text

Slide 47

Slide 47 text

Slide 48

Slide 48 text

Slide 49

Slide 49 text

Slide 50

Slide 50 text

©2018 Wantedly, Inc. 㾎վળαΠΫϧΛ͢͹΍͘ճͨ͢Ίɼ#JH2VFSZΛϑϧ׆༻ͨ͠վળͷਪਐ wϝϞϦϕʔεͷڠௐϑΟϧλϦϯά͘Β͍ͳΒͱΓ͋͑ͣΫΤϦΛॻ͍ͯ͠·͏ 㾎ͦͷଞͷ෼ੳ౳ʹ΋#JH2VFSZΛϑϧ׆༻ w#*πʔϧΛ࢖ͬͨ,1*ͳͲͷࢦඪͷ؅ཧɾ؂ࢹ w"#ςετͳͲͷΦϯϥΠϯςετͷධՁ ৴པ۠ؒɼQWBMVFɼʜ BigQuery in Recommendation Team BigQuery ͷΫΤϦΛॻ͘͜ͱʹ͸͚ͬ͜͏׳Ε͍ͯΔ(ͭ΋Γ)

Slide 51

Slide 51 text

Slide 52

Slide 52 text

Slide 53

Slide 53 text

Slide 54

Slide 54 text

Slide 55

Slide 55 text

Slide 56

Slide 56 text

Slide 57

Slide 57 text

Slide 58

Slide 58 text

Slide 59

Slide 59 text

©2018 Wantedly, Inc. Problem Definition આ໌ม਺ɿ2೔લʙ28೔લͷظؒʹ͓͚ΔɼϢʔβͷ༷ʑͳΞΫγϣϯ਺ ʲաڈ 1 ϲ݄ͷΞΫςΟϒϢʔβͷதͰɼ1 ೔લʹPush ௨஌ Λ։෧ͨ͠ϢʔβΛ౰ͯΔʳ ໨తม਺ɿ1೔લͷ Push ௨஌Λ։͍͔ͨͲ͏͔ (1 or 0) • ։͍ͨ Push ௨஌਺ • χϡʔεهࣄͷӾཡ਺ • ໊ࢗͷεΩϟϯ਺ …etc.

Slide 60

Slide 60 text

©2018 Wantedly, Inc. Problem Definition આ໌ม਺ɿ2೔લʙ28೔લͷظؒʹ͓͚ΔɼϢʔβͷ༷ʑͳΞΫγϣϯ਺ ໨తม਺ɿ1೔લͷ Push ௨஌Λ։͍͔ͨͲ͏͔ (1 or 0) • ։͍ͨ Push ௨஌਺ • χϡʔεهࣄͷӾཡ਺ • ໊ࢗͷεΩϟϯ਺ …etc. આ໌ม਺͔Β໨తม਺ (label) ͕0͔1͔Λ౰ͯΔ ʲաڈ 1 ϲ݄ͷΞΫςΟϒϢʔβͷதͰɼ1 ೔લʹPush ௨஌ Λ։෧ͨ͠ϢʔβΛ౰ͯΔʳ

Slide 61

Slide 61 text

©2018 Wantedly, Inc. Overview of architecture BigQuery merge Scheduler ʹొ࿥ PREDICT ΫΤϦͷ࣮ߦ ֘౰Ϣʔβʹ push ௨஌Λૹ৴ SELECT user_id FROM `push_predict_results` WHERE prob > #{PROB_THRESHOLD} Predict ͷ݁ՌΛ BQ ʹετΞ CREATE MODEL ΫΤϦΛ࣮ߦ ࡞੒ͨ͠Ϟσϧ Λ BQ ʹอଘ ֘౰Ϣʔβͷ id ΛಡΈࠐΉ PREDICT ΫΤϦΛ࣮ߦ͢Δ Job هड़(΄΅ SQL)

Slide 62

Slide 62 text

©2018 Wantedly, Inc. MODEL ͷ࡞੒ BigQuery merge Scheduler ʹొ࿥ PREDICT ΫΤϦͷ࣮ߦ ֘౰Ϣʔβʹ push ௨஌Λૹ৴ SELECT user_id FROM `push_predict_results` WHERE prob > #{PROB_THRESHOLD} Predict ͷ݁ՌΛ BQ ʹετΞ CREATE MODEL ΫΤϦΛ࣮ߦ ࡞੒ͨ͠Ϟσϧ Λ BQ ʹอଘ ֘౰Ϣʔβͷ id ΛಡΈࠐΉ PREDICT ΫΤϦΛ࣮ߦ͢Δ Job هड़(΄΅ SQL)

Slide 63

Slide 63 text

Slide 64

Slide 64 text

©2018 Wantedly, Inc. CREATE MODEL `ml_models.push_open` OPTIONS (model_type='logistic_reg') AS SELECT IF(open IS NULL, 0, 1) AS label, weekly_open_count, timeline_show_count, scan_count, ... FROM `people.ironna_data*` WHERE ... # Ϟσϧ࡞੒Λએݴ ˌ࡞੒͢ΔϞσϧ໊Λهड़ # ଞʹ͸ # CREATE MODEL IF NOT EXISTS # CREATE OR REPLACE MODEL CREATE MODEL `ml_models.push_open` MODEL ͷ࡞੒ MODEL ͷ࡞੒

Slide 65

Slide 65 text

©2018 Wantedly, Inc. CREATE MODEL `ml_models.push_open` OPTIONS (model_type='logistic_reg') AS SELECT IF(open IS NULL, 0, 1) AS label, weekly_open_count, timeline_show_count, scan_count, ... FROM `people.ironna_data*` WHERE ... # ௥ՃͷϞσϧΦϓγϣϯΛࢦఆ # model_type ʹ͸ # ’logistic_reg’(ϩδεςΟοΫճؼ) ͔ ‘linear_reg’(ઢܗճؼ) Λࢦఆ # ଞʹ΋ɼL1, L2 ਖ਼ଇԽͷద༻ྔ΍ɼֶशͷεςοϓ਺ɼ # ೖྗσʔλΛτϨʔχϯάηοτͱධՁηοτʹ෼ׂ͢Δํ๏ɼͳͲͳͲΛࢦఆͰ͖Δ OPTIONS (model_type='logistic_reg') AS MODEL ͷ࡞੒ MODEL ͷ࡞੒

Slide 66

Slide 66 text

©2018 Wantedly, Inc. CREATE MODEL `ml_models.push_open` OPTIONS (model_type='logistic_reg') AS SELECT IF(open IS NULL, 0, 1) AS label, weekly_open_count, timeline_show_count, scan_count, ... FROM `people.ironna_data*` WHERE ... # ೖྗσʔλΛग़ྗ͢ΔΫΤϦهड़ # label ͱ͍͏໊લͷྻΛ౰ͯΔ # (ΧϥϜ໊͸ΦϓγϣϯͰࢦఆՄೳ) SELECT IF(open IS NULL, 0, 1) AS label, weekly_open_count, timeline_show_count, scan_count, ... FROM `people.ironna_data*` WHERE ... MODEL ͷ࡞੒ MODEL ͷ࡞੒

Slide 67

Slide 67 text

©2018 Wantedly, Inc. CREATE MODEL `ml_models.push_open` OPTIONS (model_type='logistic_reg') AS SELECT IF(open IS NULL, 0, 1) AS label, weekly_open_count, timeline_show_count, scan_count, ... FROM `people.ironna_data*` WHERE ... # ೖྗσʔλΛग़ྗ͢ΔΫΤϦهड़ # label ͱ͍͏໊લͷྻΛ౰ͯΔ # (ΧϥϜ໊͸ΦϓγϣϯͰࢦఆՄೳ) SELECT IF(open IS NULL, 0, 1) AS label, weekly_open_count, timeline_show_count, scan_count, ... FROM `people.ironna_data*` WHERE ... MODEL ͷ࡞੒ MODEL ͷ࡞੒

Slide 68

Slide 68 text

Slide 69

Slide 69 text

©2018 Wantedly, Inc. SELECT * FROM ML.EVALUATE(MODEL `ml_models.push_open`, ( SELECT IF(open IS NULL, 0, 1) AS label, weekly_open_count, timeline_show_count, scan_count, ... FROM `people.ironna_data*` WHERE ... ) ) MODEL ͷੑೳධՁ ※ ਺஋͸ద౰ʹ͍ͬͯ͡·͢ # ϩδεςΟοΫճؼͷ৔߹ҎԼͷΑ͏ͳग़ྗ # precision: ద߹཰ # label ͕ 1 ͱ༧ଌͨ͠΋ͷͷ͏ͪɼ࣮ࡍʹ 1 Ͱ͋Δ΋ͷͷׂ߹ # recall: ࠶ݱ཰ # label ͕ ࣮ࡍʹ 1 Ͱ͋Δ΋ͷͷ͏ͪɼ1 ͱ༧ଌ͞Εͨ΋ͷͷׂ߹ MODEL ͷੑೳධՁ

Slide 70

Slide 70 text

Slide 71

Slide 71 text

Slide 72

Slide 72 text

©2018 Wantedly, Inc. ༧ଌΫΤϦͷεέδϡʔϥ΁ͷొ࿥ BigQuery merge Scheduler ʹొ࿥ PREDICT ΫΤϦͷ࣮ߦ ֘౰Ϣʔβʹ push ௨஌Λૹ৴ SELECT user_id FROM `push_predict_results` WHERE prob > #{PROB_THRESHOLD} Predict ͷ݁ՌΛ BQ ʹετΞ CREATE MODEL ΫΤϦΛ࣮ߦ ࡞੒ͨ͠Ϟσϧ Λ BQ ʹอଘ ֘౰Ϣʔβͷ id ΛಡΈࠐΉ PREDICT ΫΤϦΛ࣮ߦ͢Δ Job هड़(΄΅ SQL)

Slide 73

Slide 73 text

Slide 74

Slide 74 text

©2018 Wantedly, Inc. SELECT * FROM ML.PREDICT(MODEL `ml_models.push_open`, ( SELECT IF(open IS NULL, 0, 1) AS label, weekly_open_count, timeline_show_count, ... ) ) ϞσϧΛ࢖༻ͯ݁͠ՌΛ༧ଌ ※ ਺஋͸ద౰ʹ͍ͬͯ͡·͢ # ϞσϧΛ࢖༻ͯ͠ɼೖྗσʔλʹରͯ͠ label ͷ஋Λ༧ଌ͢Δ # predicted_label: ༧ଌ͞Εͨ஋ # predicted_label_probs: label ͝ͱͷ༧ଌ֬཰

Slide 75

Slide 75 text

©2018 Wantedly, Inc. PREDICT ͷ࣮ߦ BigQuery merge Scheduler ʹొ࿥ PREDICT ΫΤϦͷ࣮ߦ ֘౰Ϣʔβʹ push ௨஌Λૹ৴ SELECT user_id FROM `push_predict_results` WHERE prob > #{PROB_THRESHOLD} Predict ͷ݁ՌΛ BQ ʹετΞ CREATE MODEL ΫΤϦΛ࣮ߦ ࡞੒ͨ͠Ϟσϧ Λ BQ ʹอଘ ֘౰Ϣʔβͷ id ΛಡΈࠐΉ PREDICT ΫΤϦΛ࣮ߦ͢Δ Job هड़(΄΅ SQL)

Slide 76

Slide 76 text

Slide 77

Slide 77 text

Slide 78

Slide 78 text

Slide 79

Slide 79 text

Slide 80

Slide 80 text

Slide 81

Slide 81 text

©2018 Wantedly, Inc. Problem Definition આ໌ม਺ɿ8೔લʙ28೔લͷظؒʹ͓͚ΔɼϢʔβͷ༷ʑͳΞΫγϣϯ਺ ʲաڈ 1 ϲ݄ͷΞΫςΟϒϢʔβͷதͰɼ1 ೔લʙ7 ೔લͷؒʹ ʮ࿩Λฉ͖ʹߦ͖͍ͨʯϘλϯΛԡͨ͠ϢʔβΛ౰ͯΔʳ 1೔લʙ7೔લͷظؒʹ ʮ࿩Λฉ͖ʹߦ͖͍ͨʯϘλϯΛ ԡ͔ͨ͠Ͳ͏͔ (1 or 0) • ʮ࿩Λฉ͖ʹߦ͖͍ͨʯΛԡͨ͠਺ • ձࣾͷืूͷৄࡉϖʔδΛӾཡͨ͠਺ • ձ͔ࣾΒͷεΧ΢τϝοηʔδʹฦ৴ͨ͠਺ …etc. ໨తม਺ :

Slide 82

Slide 82 text

©2018 Wantedly, Inc. https://towardsdatascience.com/when-will-stack-overflow-reply-how-to-predict-with-bigquery-553c24b546a3 ͓΋͠Ζ͍ͳͱࢥͬͨ׆༻ྫ • ճ౴ΛಘΒΕΔ֬཰ • ճ౴ΛಘΒΕΔ·Ͱͷ࣌ؒ • ѱ͍ධՁΛೖΕΒΕΔ֬཰ • ༵೔ • ࣌ؒ • ࣭໰จͷ௕͞ • ࣭໰ͷλΠτϧͷ1ޠ໨ • λΠτϧ͕ “ʁ” ͰऴΘ͍ͬͯΔ͔Ͳ͏͔ • ࢖༻ͨ͠λά • ΞΧ΢ϯτΛ࡞੒ͨ࣌͠ظ In Stack Overflow

Slide 83

Slide 83 text

Slide 84

Slide 84 text

©2018 Wantedly, Inc. ྑ͔ͬͨ͜ͱ 㾎໰୊ʹऔΓ૊ΜͰ͔Β࠷ॳͷϞσϧΛ࡞੒͢Δ·Ͱ͕ຊ౰ʹҰॠͩͬͨ w σʔλͷҠಈͳͲΛߟ͑ͳ͍͍ͯ͘ͷͬͯૉ੖Β͍͠ʂ 㾎ҙ֎ͱ͍Ζ͍ΖͰ͖Δ w ਖ਼ଇԽɼֶश཰ͷઃఆɼFBSMZTUPQQJOHɼΫϥεෆۉߧ΁ͷରॲʜ 㾎݁Ռͷڞ༗ָ͕ w ࣮ߦ݁ՌͷϦϯΫ౤͛ͨΓɼςʔϒϧͷ৔ॴڭ͑ͨΓɼΫΤϦ͚ͩ౤͛ͨΓ 㾎ແݶͷϦιʔε w (PPHMFઌੜʹ͓෍ࢪΛ͢Δ͚ͩʜ

Slide 85

Slide 85 text

©2018 Wantedly, Inc. ؾΛ͚ͭͳ͍ͱ͍͚ͳ͍͜ͱ 㾎Ϟσϧ΍ΫΤϦͷ؅ཧ w Ϟσϧ࡞੒࣌ͷΫΤϦΛͲ͏؅ཧ͢Δʁ w ϛεͬͯϞσϧ͕ফ͞ΕͨΓ্ॻ͖͞ΕΔͱਏ͍ 㾎Ϟσϧͷߋ৽΍ఆظతͳ1SFEJDU͸΍ΓͮΒ͍ w 8BOUFEMZͰ͸͍͍ײ͡ͷεέδϡʔϥͷ࢓૊Έ͕͋Δ͕ʜ 㾎αʔϏεͰ׆༻͢Δ৔߹͸ɼऔΓ૊Ή໰୊Λ͖ͪΜͱߟ͑Δ w ͲΜͳ໰୊ʹͰ΋ద͍ͯ͠ΔΘ͚Ͱ͸ͳ͍ w બ୒ࢶͷͭͱͯ͠ߟ͑Δ

Slide 86

Slide 86 text

©2018 Wantedly, Inc. ؾΛ͚ͭͳ͍ͱ͍͚ͳ͍͜ͱ 㾎Ϟσϧ΍ΫΤϦͷ؅ཧ w Ϟσϧ࡞੒࣌ͷΫΤϦΛͲ͏؅ཧ͢Δʁ w ϛεͬͯϞσϧ͕ফ͞ΕͨΓ্ॻ͖͞ΕΔͱਏ͍ 㾎Ϟσϧͷߋ৽΍ఆظతͳ1SFEJDU͸΍ΓͮΒ͍ w 8BOUFEMZͰ͸͍͍ײ͡ͷεέδϡʔϥͷ࢓૊Έ͕͋Δ͕ʜ w ΫΤϦεέδϡʔϦϯάػೳ͕ϦϦʔε͞Εͨʂ 㾎αʔϏεͰ׆༻͢Δ৔߹͸ɼऔΓ૊Ή໰୊Λ͖ͪΜͱߟ͑Δ w ͲΜͳ໰୊ʹͰ΋ద͍ͯ͠ΔΘ͚Ͱ͸ͳ͍ w બ୒ࢶͷͭͱͯ͠ߟ͑Δ

Slide 87

Slide 87 text

©2018 Wantedly, Inc. ͜͏͍͏;͏ʹ࢖͑͹͍͍ͷ͔΋ʁ ✓ ػցֶश΍ͬͨ͜ͱͳ͍ਓ͕ػցֶशΛഽͰײ͡Δ ✓ ຊ֨తͳػցֶशϓϩδΣΫτΛ։࢝͢ΔલʹઢܗճؼͰͬ͘͞ͱ෼ੳͯ͠ΈΔ ✓ ·ͩखΛ෇͚͍ͯͳ͍͚Ͳɼ͕ͬͭΓϦιʔεΛ౤ೖ͢Δͷ͸೉͍͠Α͏ͳ୯७ ͳ෼ྨ໰୊ΛϩδεςΟοΫճؼͰͬ͘͞ͱղ͘ ✓ ΦʔϓϯʹͰ͖ΔσʔλͰ͓΋ͪΌΛ࡞ͬͯެ։͢Δ https://towardsdatascience.com/when-will-stack-overflow-reply-how-to-predict-with-bigquery-553c24b546a3

Slide 88

Slide 88 text

Slide 89

Slide 89 text

Slide 90

Slide 90 text

©2018 Wantedly, Inc. Summary 㾎 #JH2VFSZ͚ͩͰ.-͕Ͱ͖Δ#JH2VFSZ.-ͷొ৔ w σʔλͷҠಈͳ͠ w 42-Λॻ͚ͩ͘Ͱػցֶश͕Ͱ͖Δ w ػցֶशͷຽओԽɼ։ൃ଎౓ͷ޲্ 㾎 8BOUFEMZ಺Ͱ࢖ͬͯΈͨ w γϯϓϧͳ໰୊ղܾ w ຊ֨తͳ.-ϓϩδΣΫτΛ։࢝͢Δલͷௐࠪ σʔλΛ࢖͑Δ͔ͷௐࠪ 㾎 ࠓޙʹظ଴✨ ϥϯμϜϑΥϨετ͕࣮૷͞ΕͯΔʜ

Slide 91

Slide 91 text