Upgrade to Pro — share decks privately, control downloads, hide ads and more …

keisu_special_lecture_20210511.pdf

 keisu_special_lecture_20210511.pdf

05f8401462c2d4694f006634e970d577?s=128

Taro Takaguchi

May 10, 2021
Tweet

Transcript

  1. ͋Δࣄۀձࣾʹ͓͚Δ σʔλαΠΤϯεͷ࣮຿ ߴޱ ଠ࿕ LINEגࣜձࣾ Data Science ηϯλʔ ܭ਺޻ֶಛผߨٛ ౦ژେֶ

    ޻ֶ෦ ܭ਺޻ֶՊ 2021/05/11 1
  2. ߴޱ ଠ࿕ʢ͔͙ͨͪ ͨΖ͏ʣ LINEגࣜձࣾ Data Science ηϯλʔ γχΞσʔλαΠΤϯςΟετ / Ϛωʔδϟʔ

    ~2013ɹ౦ژେֶେֶӃ ৘ใཧ޻ֶܥݚڀՊ ਺ཧ৘ใֶઐ߈ ɹɹɹ ത࢜՝ఔʢ਺ཧ৘ใୈ̐ݚڀࣨʣ ~2017ɹࠃཱݚڀػؔʹͯϙευΫݚڀһɹ ܦྺ ౰࣌ͷઐ໳෼໺ 2 ωοτϫʔΫՊֶʢಛʹ࣌ؒతʹมԽ͢ΔωοτϫʔΫʣ
  3. اۀʹస͖͔͚ͨͬ͡ 3 2 2 2 1 1 1 3 3

    4 4 3 4 ໾ʹཱͪͦ͏ɺͦΕͰ΋࣮ࣾձͱͷڑ཭͸ԕ͍… @ LINE DEVELOPER DAY 2019 σʔλΛ׆༻ͨ͠ࣄۀͷ࠷લઢΛ ݟ͍ͨɾؔΘΓ͍ͨ
  4. ςʔϚ ʮ͋Δࣄۀձࣾʹॴଐ͢ΔσʔλαΠΤϯςΟετ͕ɺ
 ͲΜͳ࢓ࣄͰษڧ΍ݚڀͷܦݧΛ׆͔͍ͯ͠Δ͔ʁʯ 4

  5. શମͷߏ੒ 1. ର৅Λ஌Δɿ
 ࣄۀձࣾͷσʔλαΠΤϯςΟετͬͯͲΜͳ࢓ࣄʁ ʢٳܜʣ 2. த਎Λ஌Δɿ
 ࣮຿ͰΑ͘༻͍Δ౷ܭͷҰ෦ͱ۩ମతͳࣄྫ 5

  6. 1. ର৅Λ஌Δɿ
 ࣄۀձࣾͷσʔλαΠΤϯςΟετͬͯ
 ͲΜͳ࢓ࣄʁ 6

  7. ͦ΋ͦ΋σʔλαΠΤϯςΟετͱ͸ʁ 7 - اۀɾ࣌ظɾίϛϡχςΟʹΑΓఆٛ͸༷ʑ - ಉ͡৬໊ͰҟͳΔۀ຿ɺҟͳΔ৬໊Ͱڞ௨͢Δۀ຿ ࢦඪΛఆٛ͠ܭଌ͢Δ / ετʔϦʔΛޠΔ /

    πʔϧΛ࡞Δ Analyticsʢ෼ੳܕʣ ػցֶशͷख๏Λ੡඼ɾαʔϏεʹ࣮૷͢Δ AlgorithmsʢΞϧΰϦζϜܕʣ ౷ܭख๏ʹΑΓҼՌؔ܎Λཱূ͢Δ Inferenceʢਪ࿦ܕʣ Ref. https://www.linkedin.com/pulse/one-data-science-job-doesnt-fit-all-elena-grewal/ Data Scientist ෼ྨͷҰྫɿ
  8. λεΫ͝ͱʹ໾ׂ͕෼͔ΕΔ 8 Q1. ੡඼ɾαʔϏεʹ࣮૷͞ΕΔίʔυΛॻ͘ʁ Analytics ʢ෼ੳܕʣ Algorithms ʢΞϧΰϦζϜܕʣ Inference ʢਪ࿦ܕʣ

    Q2. ౷ܭख๏ʹΑΓҼՌؔ܎Λݕূ͢Δʁ Yes Yes No No ෼ྨͷҰྫɿ
  9. ૊৫ߏ੒΋୲౰ྖҬʹରԠ͍ͯ͠Δ 9 Data Science ηϯλʔ Data Science Machine Learning Machine

    Learning
 Research Analyticsʢ෼ੳܕʣ AlgorithmsʢΞϧΰϦζϜܕʣ Inferenceʢਪ࿦ܕʣ جૅݚڀ͓Αͼࣄۀ΁ͷԠ༻ ػցֶशΤϯδχΞ
  10. ෼ੳɾਪ࿦ܕͷ۩ମతͳ࢓ࣄ ਐߦத ࣄલ ࣄޙ ࣌ظ 10 Ωϟϯϖʔϯ / ৽ػೳͷ௥Ճ /

    طଘػೳͷมߋͳͲ
  11. ෼ੳɾਪ࿦ܕͷ۩ମతͳ࢓ࣄ ਐߦத ࣄલ ࣄޙ ࣌ظ 11 Ωϟϯϖʔϯ / ৽ػೳͷ௥Ճ /

    طଘػೳͷมߋͳͲ - Ωϟϯϖʔϯͷ৚݅બఆ - ৽ػೳͷχʔζݟੵ΋Γ - ػೳมߋͷӨڹͷݟੵ΋Γ - etc.
  12. ෼ੳɾਪ࿦ܕͷ۩ମతͳ࢓ࣄ ਐߦத ࣄલ ࣄޙ ࣌ظ 12 Ωϟϯϖʔϯ / ৽ػೳͷ௥Ճ /

    طଘػೳͷมߋͳͲ - ΦϯϥΠϯ A/B ςετ - μογϡϘʔυͷ࡞੒
 ओཁͳࣄۀࢦඪͷϞχλϦϯάද - ҟৗͳมԽͷݕग़ - etc.
  13. ෼ੳɾਪ࿦ܕͷ۩ମతͳ࢓ࣄ ਐߦத ࣄલ ࣄޙ ࣌ظ 13 Ωϟϯϖʔϯ / ৽ػೳͷ௥Ճ /

    طଘػೳͷมߋͳͲ - ࢪࡦͷޮՌݕূ - ҼՌਪ࿦ - ௕ظతมԽͷཁҼ෼ղ - etc.
  14. νʔϜɺϓϩδΣΫτɺϓϩμΫτ 14 νʔϜ

  15. νʔϜɺϓϩδΣΫτɺϓϩμΫτ 15 νʔϜ ϓϩδΣΫτ

  16. νʔϜɺϓϩδΣΫτɺϓϩμΫτ 16 νʔϜ ϓϩδΣΫτ ϓϩμΫτ ྫɿϓϩμΫτʮLINE ΞϓϦʯͷ
 ◦◦ػೳ௥ՃϓϩδΣΫτʹؔΘΔ Data Science

    νʔϜ
  17. ʢνʔϜ|ϓϩδΣΫτ|ϓϩμΫτʣϚωδϝϯτ 17 νʔϜ ϓϩδΣΫτ ϓϩμΫτ ૊৫ͷ໨ඪΛઃఆ͠ɺͦͷ࣮ݱͷͨΊʹ ࿑ྗɾ࣌ؒɾ͓ۚͷ഑෼Λௐ੔͠ޮ཰Խ͢Δ ※ આ໌ͷͨΊʹ୯७Խ͍ͯ͠·͢

  18. ʮయܕతͳ̍೔ͷ࢓ࣄ಺༰͸ʁʯ 18 ࣌ظ ϓϩδΣΫτ A ϓϩδΣΫτ B ϓϩδΣΫτ C ͱ͋Δ

    1೔ λεΫ͕ؒΛۭ͚ͯஅଓతʹਐߦ͢Δ e.g. ଞνʔϜͷਐߦ଴ͪɺಥൃతͳґཔ
  19. ෼ੳɾਪ࿦ܕͷλεΫɿ՝୊ղܾͷαΠΫϧ ؍ଌ Ծઆͱ՝୊ ͷઃఆ ݕূ ղܾࡦͷཱҊ 19 ࣌ظ

  20. εςοϓ̍. ؍ଌ ؍ଌ Ծઆͱ՝୊ ͷઃఆ ݕূ ղܾࡦͷཱҊ 20 ࣌ظ ࠷ۙɺΞΫςΟϒϢʔβʔ਺


    ͕ఀ଺͍ͯ͠Δʁ 2݄ 3݄ 4݄ 5݄ μογϡϘʔυɿ ओཁͳࣄۀࢦඪͷϞχλϦϯάද ※ ΍ΓऔΓͱ਺஋͸͢΂ͯՍۭͷ΋ͷ
  21. εςοϓ̎. Ծઆͱ՝୊ͷઃఆ ؍ଌ Ծઆͱ՝୊ ͷઃఆ ݕূ ղܾࡦͷཱҊ 21 ࣌ظ ͜ͷΞΫςΟϒϢʔβʔ਺ͷ


    ਪҠ͸ରॲ͢΂͖΋ͷ͔ʁ - ྫ೥ͷقઅతͳมಈʁ - Ϣʔβʔͷηάϝϯτ͝ͱͷมԽʁ - ৽ن / طଘ / ෮ؼ - ଞػೳͷར༻ϢʔβʔͷਪҠʁ → ʮ৽نϢʔβʔͷܧଓ཰͕௿Լ
 ͍ͯ͠Δɻݩͷਫ४ʹճ෮͢Δͱ
 ˓ສਓ૿ՃͷӨڹ͕͋Δʯ ※ ΍ΓऔΓͱ਺஋͸͢΂ͯՍۭͷ΋ͷ
  22. εςοϓ̏. ղܾࡦͷཱҊ ؍ଌ Ծઆͱ՝୊ ͷઃఆ ݕূ ղܾࡦͷཱҊ 22 ࣌ظ -

    ৽نϢʔβʔʹϩάΠϯΛ ଅ͢௨஌ΛૹΖ͏ - ௨஌ͷස౓Λςετ͍ͨ͠ ςετͷઃܭΛ͠·͢ - ੒൱ΛධՁ͢Δࢦඪͷܾఆ - ςετʹඞཁͳαϯϓϧ
 αΠζͷܭࢉ - ੒൱ͷ൑அج४ͷ߹ҙ ※ ΍ΓऔΓͱ਺஋͸͢΂ͯՍۭͷ΋ͷ
  23. εςοϓ̐. ݕূ ؍ଌ Ծઆͱ՝୊ ͷઃఆ ݕূ ղܾࡦͷཱҊ 23 ࣌ظ ςετͷ݁ՌΛ෼ੳ͠·͢

    - σʔλͷਖ਼ৗͳऩूͷ֬ೝ - ࢦඪʹର͢ΔԾઆݕఆ - ௥ՃͷվળҊͷࣔࠦ - ૯߹తͳϨϙʔςΟϯά Ճೖཌ೔ʹ̍ճ͚ͩ௨஌Λ
 ૹΔҊΛ࠾༻͢Δ ※ ΍ΓऔΓͱ਺஋͸͢΂ͯՍۭͷ΋ͷ
  24. εςοϓ̍(2). ؍ଌ ؍ଌ Ծઆͱ՝୊ ͷઃఆ ݕূ ղܾࡦͷཱҊ 24 ࣌ظ ৽نϢʔβʔͷܧଓ཰͸


    ࠓޙ΋ϞχλϦϯά͠·͢ ※ ΍ΓऔΓͱ਺஋͸͢΂ͯՍۭͷ΋ͷ 2݄ 3݄ 4݄ 5݄ 6݄ 2݄ 3݄ 4݄ 5݄ 6݄ μογϡϘʔυʹ߲໨Λ௥Ճ͢Δ ΞΫςΟϒϢʔβʔ਺ ৽نϢʔβʔܧଓ཰
  25. ෼ੳɾਪ࿦ܕͷλεΫɿ՝୊ղܾͷαΠΫϧ ؍ଌ Ծઆͱ՝୊ ͷઃఆ ݕূ ղܾࡦͷཱҊ 25 ࣌ظ - ෼ੳɾਪ࿦ͷλεΫ͸


    ؔ܎ऀͱͷίϛϡχέʔγϣϯΛ ௨ͯ͡ਐߦ͢Δ - ౷ܭͳͲઐ໳஌ࣝͷ׆༻͸ɺ
 શମͷαΠΫϧͷதͷҰཁૉ - ࠷ऴతͳҙࢥܾఆऀ͸ɺࣄۀɾ
 ϓϩμΫτɾϓϩδΣΫτͷ੹೚ऀ
  26. ઐ໳తͳ਺ֶͷ஌ࣝ͸࢖͏ʁ 26 “LIFE AND MATHS”, © Pearls of Raw Nerdism

    http://pearlsofrawnerdism.com/life-and-maths/
  27. ઐ໳తͳ਺ֶͷ஌ࣝ͸࢖͏ʁ 27 “LIFE AND MATHS”, © Pearls of Raw Nerdism

    http://pearlsofrawnerdism.com/life-and-maths/ ࢲͷߟ͑ɿ - ઐ໳తͳ਺ֶͳ͠Ͱ΋ࡁΉػձͷ΄͏͕ଟ͍ - ઐ໳஌͕ࣝ͋Δͱɺ՝୊ղܾͷ֤εςοϓͷ্࣭͕͕Δ
  28. ઐ໳తͳ਺ֶͳ͠Ͱ΋ࡁΉػձͷ΄͏͕ଟ͍ ൃੜස౓ ਺ֶతͳ ෳࡶ౓ 28 ߴ ௿ ߴ ௿ ֓೦ਤ

    ෼ੳɾਪ࿦ܕͷ໾ׂ ˚ෳࡶͳ͜ͱΛ਱ߦ͢Δ͜ͱ ˚ཧ࿦తʹ৽نͳ͜ͱΛߦ͏͜ͱ ✓ ࣄۀʹ໾ཱͭ஌ݟΛద੾ʹఏڙ͢Δ͜ͱ ʮࣄۀʹର͢Δߩݙ౓ʯ ʮ࣮ࢪʹཁ͢Δίετʯͷ͕࣠ӅΕ͍ͯΔ ਺ֶతͳ೉͠͞ ≠ ࣄۀ্ͷ՝୊ղܾͷ೉͠͞
  29. ฏқͳ࡞ۀ͸ɺઐ໳తͳۀ຿ͷ౔୆ ྫɿ୯७ͳूܭ࡞ۀ ઐ໳஌ࣝΛ
 ൃش͢Δۀ຿ 29 ฏқͳ࡞ۀΛ௨ͨ͡σʔλɾࣄۀ΁ͷशख़ → ઐ໳஌ࣝΛൃش͢Δۀ຿ͷ਱ߦ ࣄۀʹର͢Δཧղ౓ͷ޲্ →

    ࣮ࢪʹίετͷ͔͔Δ෼ੳλεΫͷड೚
  30. ઐ໳஌͕ࣝ͋Δͱɺ՝୊ղܾͷ֤εςοϓͷ্࣭͕͕Δ 30 ؍ଌ Ծઆͱ՝୊ͷઃఆ ݕূ ղܾࡦͷཱҊ ΑΓϊΠζʹؤ݈Ͱղऍͷ͠΍͍͢ࢦඪΛ ఆٛͰ͖Δ ʮσʔλͱ਺ֶʹΑͬͯղ͚Δ໰୊ʯͷ ఆࣜԽͷϨύʔτϦʔ͕૿͑Δ

    ద੾͔ͭޮ཰తͳղܾࡦΛબ୒Ͱ͖Δ ҙࢥܾఆʹඞཁͳ஌ݟΛత֬ʹநग़Ͱ͖Δ
  31. ֓೦ͷ֫ಘ͸ੈքͷݟ͑ํΛม͑Δ 31 ՝୊ɿ̎Λ̍ສݸ଍ͨ͠౴͑Λ஌Γ͍ͨ ৐ࢉͷ֓೦Λ஌Βͳ͍ͱ 2 + 2 + 2 +

    2 + …… ʮݱ࣮తͳ࣌ؒͰ͸ղܾͰ͖·ͤΜʯ ৐ࢉΛ஌͍ͬͯΕ͹ 2 × 10,000 = 20,000 ղ͚ͳ͍໰୊ ղ͚Δ໰୊
  32. ʮ౴͑Λग़͢ͱࣄۀʹ໾ཱͭʯྖҬΛ໨ࢦ͢ 32 ࣄۀՁ஋ʹ ݁ͼͭ͘ ࣄۀՁ஋ʹ ݁ͼ͖ͭͮΒ͍ ౴͑Λग़ͤΔ ౴͑Λग़ͤͳ͍ ઐ໳஌ࣝͷशಘ ࣄۀͷཧղ

    ؔ܎ऀͱͷର࿩ Cf. ҆୐࿨ਓ, ʮΠγϡʔ͔Β͸͡ΊΑʕ஌తੜ࢈ͷʰγϯϓϧͳຊ࣭ʱʯ, ӳ࣏ग़൛ʢ2010ʣ σʔλαΠΤϯςΟετͷۀ຿্ͷλεΫΛ̎࣍ݩʹϚοϓ͢Δ
  33. ࣄۀձࣾͷσʔλαΠΤϯςΟετͷ࢓ࣄ 33 ࣄۀͷͨΊͷ՝୊ղܾͷαΠΫϧ ෼ੳܕ / ΞϧΰϦζϜܕ / ਪ࿦ܕ νʔϜͱͯ͠ϓϩδΣΫτʹऔΓ૊Ή ෼ྨʢҰྫʣ

    Ґஔ͚ͮ ෼ੳɾਪ࿦ͷλεΫ ઐ໳తͳ஌ࣝ ՝୊ղܾͷ࣭Λ্͛Δ
  34. ͲΜͳ؀ڥͩͱྗΛൃش͠΍͍͔͢ʁ 34 A. αΠΤϯε͕Ͱ͖Δ͜ͱ Պֶతํ๏ʹج͍ͮͯۀ຿Λ਱ߦ͠ɺ੒Ռ͕ೝ஌͞ΕΔ͜ͱ - ٬؍తͳࠜڌʹج͍ͮͯɺ࿦ཧΛల։͢Δ͜ͱ - खଓ͖͕ه࿥͞Εɺ࠶ݱՄೳͰ͋Δ͜ͱ -

    ͱ͘ʹ݁࿦͕ޡΓͩͬͨ৔߹ʹɺݕূՄೳͰ͋Δ͜ͱ
  35. ۀ຿ΛαΠΤϯεʹ͢ΔͨΊʹ 1. ܧଓ͢Δ 2. ԾఆΛڞ༗͢Δ 3. ਺ࣈΛݟΔલʹ൑அج४ΛܾΊΔ 35 σʔλαΠΤϯςΟετଆʹ΋৺͕͚Δ΂͖͜ͱ͕͋Δ

  36. 1. ܧଓ͢Δ 36 Պֶతํ๏͸ɺ܁Γฦ͢͜ͱʹҙ͕ٛ͋Δ ԿΛ͢΂͖͔ʁ - ࠶ݱɾݕূՄೳͳΑ͏ʹه࿥Λ࢒͢ - ҡ࣋Մೳͳ؍ଌํ๏Λߏங͢Δ
 ʢϞχλϦϯάͷࣗಈԽʣ

    - ࣍ͷ՝୊ઃఆΛଅࣔࠦ͢Λఏڙ͢Δ
  37. 2. ԾఆΛڞ༗͢Δ 37 ܦݧՊֶʹ͓͚ΔՊֶత஌ࣝ͸ ✗ ઈରෆมͷਅ࣮ͷू߹ ✓ ؍ଌͱԾఆʹج͍ͮͯਪ࿦͞Εͨؼ݁ ԿΛ͢΂͖͔ʁ -

    ԾఆΛ໌֬ʹ఻͑Δ ʮϢʔβʔ਺ͷ૿Ճ཰͸ઌ݄ͱಉ͡ͱԾఆ͠·͢ʯ - ݕূͷεςοϓͰ͸ɺࣄલͷԾఆͷଥ౰ੑ΋ݕূ͢Δ ʮϢʔβʔ਺ͷ૿Ճ཰͸ɺ݁Ռతʹઌ݄ͱൺ΂ͯʙͰͨ͠ʯ
  38. 3. ਺ࣈΛݟΔલʹ൑அج४ΛܾΊΔ 38 ਺ྔ → ೔ৗݴޠͷม׵ʹ͸ᐆດੑ͕͋Δ ͜ͷࢦඪ͕ “े෼ʹ” ্ঢͨ͠Β
 ςετ͸੒ޭͱ൑அ͠·͠ΐ͏

    ʢ+3% ͸ ”े෼” ͩΖ͏͔…ʣ ԿΛ͢΂͖͔ʁ - ࣄલʹ൑அج४ΛྔతʹܾΊΔ - ج४ͷࠜڌ͸٬؍తʹ͢Δ
 (ྫ) ࣄۀ໨ඪʹର͢Δظ଴د༩
 ɹɹ౤͡ΒΕͨίετͷճऩ
 ɹɹաڈͷྨࣅࣄྫͷ݁Ռ ※ ਺஋͸͢΂ͯՍۭͷ΋ͷ ࢦඪͷ্ঢ͸ +3% Ͱͨ͠
  39. ٳܜ 39

  40. શମͷߏ੒ 1. ର৅Λ஌Δɿ
 ࣄۀձࣾͷσʔλαΠΤϯςΟετͬͯͲΜͳ࢓ࣄʁ ʢٳܜʣ 2. த਎Λ஌Δɿ
 ࣮຿ͰΑ͘༻͍Δ౷ܭͷҰ෦ͱ۩ମతͳࣄྫ 40

  41. 2. த਎Λ஌Δɿ
 ࣮຿ͰΑ͘༻͍Δ౷ܭͷҰ෦ͱ۩ମతͳࣄྫ 41

  42. ෼ੳɾਪ࿦ͷ۩ମతͳ࢓ࣄʢ࠶ܝʣ 42 ਐߦத ࣄલ ࣄޙ ࣌ظ Ωϟϯϖʔϯ / ৽ػೳͷ௥Ճ /

    طଘػೳͷมߋͳͲ ΦϯϥΠϯ A/B ςετ 1. αϯϓϧαΠζͷܭࢉ 2. ଟॏൺֱ 3. ׳ΕޮՌͷਪఆ
  43. 1. αϯϓϧαΠζͷܭࢉ 43 ςετର৅ͷࠩΛݕఆ͢ΔͨΊʹඞཁͳαϯϓϧαΠζΛࢉग़͢Δ͜ͱ Q. αϯϓϧαΠζΛܭࢉܾͯ͠ΊΔཧ༝͸ʁ A. ఻౷తͳԠ༻෼໺Ͱ͸ɺαϯϓϧऩूͷίετ͕ߴ͔ͬͨ
 ɹe.g. ྟচݚڀɺ೶ۀ

    Q. ΢ΣϒαʔϏεͳΒαϯϓϧऩूͷίετ͸ߴ͘ͳ͍ͷͰ͸ʁ
  44. ΢ΣϒαʔϏεͰ΋αϯϓϧαΠζΛܭࢉ͢Δཧ༝ 44 1. ա৒ʹେ͖ͳαϯϓϧαΠζ → খ͞ͳมԽͰ΋༗ҙʹͳΓ͕ͪ ʮ౷ܭతʹ༗ҙʯ͸ڧ͍ҹ৅Λ༩͑Δදݱ 2. ಛʹςετҊ͕ྑ͘ͳ͍࣌ɺϢʔβʔʹແ༻ͳӨڹΛ༩͑ͯ͠·͏ 4.

    P-Hacking ͷ༨஍͕࢒Δ ʮ༗ҙ͕ࠩग़ͳ͔͔ͬͨΒαϯϓϧαΠζΛେ͖ͯ͘͠࠶ςετ͠Α͏ʯ 3. SUTVA (Stable Unit Treatment Value Assumption) ͕ഁΕ΍͘͢ͳΔ ʮ͋ΔϢʔβʔͷߦಈ͸ଞͷϢʔβʔͷׂΓ౰ͯʹӨڹ͞Εͳ͍ʯ ʢྫʣςετը໘͕ڞ༗͞ΕΔɺϝσΟΞʹऔΓ্͛ΒΕΔ
  45. αϯϓϧαΠζܭࢉͷجຊܗ 45 ઃఆ - ಠཱͳ̎܈αϯϓϧͷฏۉͷݕఆ - ฼෼ࢄ͸̎܈Ͱಉ͡ & ط஌ -

    αϯϓϧαΠζ͸਺ઍ ~ ਺ສ݅ఔ౓͸औΕΔ ݕఆ͞ΕΔԾઆ - ؼແԾઆ - ରཱԾઆ H0 H1 μ1 − μ2 = 0 μ1 − μ2 ≠ 0 αϯϓϧαΠζɹͷܾఆʹඞཁͳύϥϝʔλ - ༗ҙਫ४ - ݕग़ྗ - ޮՌྔ - ฼෼ࢄ α 1 − β δ = μ1 − μ2 σ2 < + ∞ n ʢɹ ͕ਅͷ৔߹ʣ H1
  46. αϯϓϧαΠζܭࢉͷ෮शʢ̍ʣ 46 ਤ͸ԼهจݙΑΓ࠶ߏ੒ͨ͠ Gerald van Belle, “Statistical Rules of Thumb”

    (2nd edition), Wiley, 2008 ඪຊฏۉͷࠩ x1 − x2 H0 : μ1 − μ2 = 0 0 S . E . = σ 2 n ਖ਼ن෼෍ͷ࠶ੜੑΑΓ α 2 α 2 ༗ҙਫ४ɹɿ α ɹ͕ਅͷͱ͖ɹ Λ࠾୒ͯ͠͠·͏֬཰ ʢِཅੑʣ H0 H1
  47. αϯϓϧαΠζܭࢉͷ෮शʢ̎ʣ 47 ਤ͸ԼهจݙΑΓ࠶ߏ੒ͨ͠ Gerald van Belle, “Statistical Rules of Thumb”

    (2nd edition), Wiley, 2008 ඪຊฏۉͷࠩ x1 − x2 H1 : μ1 − μ2 = δ δ S . E . = σ 2 n H0 : μ1 − μ2 = 0 0 β = 1− ݕग़ྗ (1 − β) ɹ͕ਅͷͱ͖ɹ Λ࠾୒ͯ͠ ͠·͏֬཰ʢِӄੑʣ H0 H1 β
  48. αϯϓϧαΠζܭࢉͷ෮शʢ̏ʣ 48 ਤ͸ԼهจݙΑΓ࠶ߏ੒ͨ͠ Gerald van Belle, “Statistical Rules of Thumb”

    (2nd edition), Wiley, 2008 ඪຊฏۉͷࠩ x1 − x2 δ S . E . = σ 2 n 0 β n* = 2σ2 (z1−α/2 + z1−β) 2 δ2 㱺 ཁ݅Λຬͨͨ͢Ίʹ ࠷௿ݶඞཁͳαϯϓϧαΠζ z1−α/2 σ 2 n* = δ − z1−β σ 2 n* ඪ४ਖ਼ن෼෍ͷ ෼Ґ఺ؔ਺ α 2
  49. 1. ਅͷ෼෍ͷ෼ࢄɹ ͕େ 2. ِཅੑɺِӄੑΛ཈͑Δ
 ɹɹ͕খ 3. ݕग़͍ͨ͠ޮՌྔɹ͕খ ͕େ͖͘ͳΔཁҼ n*

    σ2 α, β δ αϯϓϧαΠζʹ͍ͭͯͷิ଍ 49 n* = 2σ2 (z1−α/2 + z1−β) 2 δ2 ཁ݅Λຬͨͨ͢Ίʹ
 ࠷௿ݶඞཁͳαϯϓϧαΠζ ύϥϝʔλͷܾΊํʢҰྫʣ ɹɹɿ׳शతͳ஋͔ɺ΍΍ݫ͠໨ʹ ɹɹɿ௚ۙͷ࣮ଌ஋
 ɹɹʢςετޙʹଥ౰ੑΛ֬ೝʣ ɹɹɿ׳शతͳ஋
 ɹɹɹor ίετΛ্ճΔޮՌ
 ɹɹɹor աڈͷྨࣅ͢Δςετ݁Ռ α, β σ2 δ
  50. 2. ଟॏൺֱ 50 ࣮຿Ͱ͸ɺ̏܈Ҏ্ͷൺֱΛٻΊΒΕΔ͜ͱ͕Α͋͘Δ 㲗 લઅͰ෮शͨ̎͠܈ؒͷݕఆ എܠ ୹ظؒͰͳΔ΂͘ଟ͘ͷՄೳੑΛςετ͍ͨ͠ ܈ؒϖΞͷճ਺͚ͩ୯७ʹݕఆΛ܁Γฦͯ͠͸͍͚ͳ͍ ʂʂ

    ଟॏൺֱ ͷ໰୊ →
  51. ࣮ྫɿάϧʔϓ࡞੒ը໘ͷมߋςετ 51 “ίϛϡχέʔγϣϯΞϓϦʮLINEʯͷػೳվળΛࢧ͑ΔσʔλαΠΤϯε” LINE DEVELOPER DAY 2019 https://linedevday.linecorp.com/jp/2019/sessions/B1-3 άϧʔϓ࡞੒ͷखॱΛɺΑΓ࢖͍΍͘͢վྑ͍ͨ͠

  52. ̎ͭͷมߋΛ૊Έ߹Θͤͯࢼ͍ͨ͠ 52 “ίϛϡχέʔγϣϯΞϓϦʮLINEʯͷػೳվળΛࢧ͑ΔσʔλαΠΤϯε” LINE DEVELOPER DAY 2019 https://linedevday.linecorp.com/jp/2019/sessions/B1-3 1. ࠷ۙτʔΫͨ͠༑ͩͪΛ༏ઌදࣔ͢Δ

    2. खॱΛ̍ը໘ʹ·ͱΊΔ
  53. ̎×̎=̐௨Γ̒ϖΞͷݕఆʁ 53 “ίϛϡχέʔγϣϯΞϓϦʮLINEʯͷػೳվળΛࢧ͑ΔσʔλαΠΤϯε” LINE DEVELOPER DAY 2019 https://linedevday.linecorp.com/jp/2019/sessions/B1-3 গ਺ͷީิͰ΋ɺৄࡉͳݕূ͸ҙ֎ͳ΄ͲෳࡶʹͳΔ

  54. ݕఆͷ܁Γฦ͠͸Կ͕໰୊͔ʁ 54 ݕఆ͞ΕΔԾઆ - ؼແԾઆ - ରཱԾઆ H0 H1 θ1

    = θ2 = θ3 = θ4 ʢ̐܈ͷ৔߹ʣ {θi} i=1,2,3,4 ͷ͏ͪগͳ͘ͱ΋̍ͭͷϖΞͰ θi ≠ θj (i ≠ j) ࣮ߦతͳ༗ҙਫ४ Family-Wise Error Rate α = 1 − (1 − α)6 ≥ α α α 1 − (1 − α)6 શମͱͯ͠ݟͨ࣌ʹɺِཅੑ཰্͕͕ͬͯ͠·͏
  55. Bonferroni ิਖ਼ 55 ֤ϖΞͷݕఆͷ༗ҙਫ४Λɺݕఆͷճ਺ɹͰׂͬͨ஋ʹௐ੔͢Δ α → α m m Family-Wise

    Error Rate α ≤ α ͱͳΓɺશମͱͯ͠ͷ༗ҙਫ४͕อͨΕΔ σϝϦοτ ͕େ͖͍ͱอकతʹͳΓ͕ͪʢِӄੑ཰ͷ্ঢʣ m
  56. άϧʔϓ࡞੒ը໘ςετͰͷରॲ 56 “ίϛϡχέʔγϣϯΞϓϦʮLINEʯͷػೳվળΛࢧ͑ΔσʔλαΠΤϯε” LINE DEVELOPER DAY 2019 https://linedevday.linecorp.com/jp/2019/sessions/B1-3 - ࣄલݕূʹج͍ͮͯɺର৅Λ̏܈̎ϖΞʹߜΔ

    - ̎ϖΞʹରͯ͠ Bonferroni ิਖ਼ͯ͠ݕఆ͢Δ α → α/2
  57. άϧʔϓ࡞੒ը໘ςετͷ݁Ռ 57 “ίϛϡχέʔγϣϯΞϓϦʮLINEʯͷػೳվળΛࢧ͑ΔσʔλαΠΤϯε” LINE DEVELOPER DAY 2019 https://linedevday.linecorp.com/jp/2019/sessions/B1-3 ʮ̎ը໘ +

    ࠷ۙτʔΫͨ͠༑ͩͪϦετʯ → ࡞੒׬ྃ཰Λҡ࣋ͭͭ͠ɺ࡞੒ͷॴཁ࣌ؒΛ୹ॖͨ͠
  58. ଟॏൺֱ΁ͷରॲʹऔΓೖΕ͍ͯΔ͜ͱ 58 - جຊతʹൺֱର৅Λߴʑ̐ύλʔϯʹݶఆ͢Δ
 ̑ύλʔϯҎ্͸෼ੳ΋ղऍ΋೉͘͠ͳΔ - Bonferroni ิਖ਼Ͱِཅੑ཰Λ཈੍͢Δ
 ϦεΫͷ͋ΔςετͰ͸ِཅੑΛආ͚͍ͨ -

    σϝϦοτΛิ͏ͨΊɺݕग़ྗɹΛߴΊʹઃఆ͢Δ β
  59. 3. ׳ΕޮՌͷਪఆ 59 ΦϯϥΠϯςετͷظؒ͸௨ৗ̎ʙ̏िؒ ୹ظؒͷ൓ԠΛͦͷ··ड͚औͬͯΑ͍ͷͩΖ͏͔ʁ Ծઆ ಛʹྺ࢙͕௕͘श׳Խ͍ͯ͠Δػೳ΄Ͳɺ ը໘ͷมߋʹର͢ΔҰ࣌తͳ൓Ԡ͕ݱΕΔ ՝୊ Ұ࣌తͳ൓Ԡ͕ఆৗతͳར༻ʹམͪண͘

    ʮ׳ΕޮՌʯΛݕग़͍ͨ͠
  60. ࣮ྫɿ༑ͩͪ௥Ճը໘ͷγϯϓϧԽςετ 60 “ίϛϡχέʔγϣϯΞϓϦʮLINEʯʹ͓͚Δ࣮ફతσʔλαΠΤϯε” DEIM 2020 https://engineering.linecorp.com/ja/blog/deim2020-report/ - ༑ͩͪ௥Ճը໘͔Βɺ༑ͩͪ௥ՃҎ֎ͷΞΠςϜΛ࡟আ͢Δ - ༑ͩͪ௥Ճ਺

    & LINEެࣜΞΧ΢ϯτ௥Ճ਺͕ݮগͯ͠͠·ͬͨ
  61. LINE ͷ৽نϢʔβʔͷΈʹݶఆͯ͠ܭଌͯ͠ΈΔ 61 - ༑ͩͪ௥Ճ਺ɾLINEެࣜΞΧ΢ϯτ௥Ճ਺ͱ΋ʹ༗ҙͳมԽͳ͠ - Ծઆɿશମʹ͸طଘϢʔβʔ͕׳ΕΔ·Ͱͷ൓Ԡ͕ݱΕ͍ͯΔʁ “ίϛϡχέʔγϣϯΞϓϦʮLINEʯʹ͓͚Δ࣮ફతσʔλαΠΤϯε” DEIM 2020

    https://engineering.linecorp.com/ja/blog/deim2020-report/
  62. ׳ΕޮՌΛࠩ෼ͷࠩͰϞσϧԽ͢Δ 62 1st half 2nd half Control Treatment yT,1 yC,1

    yC,2 yT,2 ׳ΕޮՌҎ֎ͷӨڹ͸̎܈ͰಉҰ ʢฒߦτϨϯυ & ڞ௨γϣοΫͷԾఆʣ Ծఆ ςετظؒΛલɾޙ൒ʹ̎෼͢Δ ࠩ෼ͷࠩ౷ܭྔ δ = (yT,2 − yC,2) − (yT,1 − yC,1) ճؼϞσϧԽ ̂ β3 = ̂ δ y = β0 + β1 T + β2 S + β3 TS + ε T / C ͷμϛʔ 1st / 2nd ͷμϛʔ Ͱ͋Γɺ ճؼϞσϧͷ౰ͯ͸·Γ & ܎਺ͷ༗ҙੑΛ֬ೝ͢Δ
  63. ༑ͩͪ௥Ճը໘ͷγϯϓϧԽςετͷ݁Ռ 63 - LINEެࣜΞΧ΢ϯτ௥Ճ਺ͷมԽʹ͸ɺ׳ΕޮՌ͕ݱΕ͍ͯͨ - LINEެࣜΞΧ΢ϯτ௥Ճͷ࿮ͷΈ࡟আͯ͠ɺϦϦʔε͞Εͨ “ίϛϡχέʔγϣϯΞϓϦʮLINEʯʹ͓͚Δ࣮ફతσʔλαΠΤϯε” DEIM 2020 https://engineering.linecorp.com/ja/blog/deim2020-report/

  64. ख๏Λඪ४Խͯ͠ਫฏల։͢Δ 64 “σʔλαΠΤϯε͕ಋ͘τʔΫϝχϡʔUIͷϦχϡʔΞϧϓϩδΣΫτ” LINE DEVELOPER DAY 2020 https://linedevday.linecorp.com/2020/ja/sessions/3932 - ׳ΕޮՌݕग़๏͸ɺτʔΫϝχϡʔͷϦχϡʔΞϧͰ΋׆༻ͨ͠

    - ۀ຿ΛαΠΤϯεʹ͢ΔͨΊʹ - ܧଓ͢Δ
  65. ࣮຿ͰΑ͘༻͍Δ౷ܭͷҰ෦ͱ۩ମతͳࣄྫ 65 ΦϯϥΠϯ A/B ςετ 1. αϯϓϧαΠζͷܭࢉ 2. ଟॏൺֱ 3.

    ׳ΕޮՌͷਪఆ
  66. ͓ΘΓʹ 66

  67. ʮσʔλαΠΤϯςΟετʯͷকདྷʁ 67 ࢲͷߟ͑ɿ - ʮσʔλαΠΤϯςΟετʯͱ͍͏ݺশ΍ظ଴͸มΘ͍͔ͬͯ͘΋ - σʔλͱ਺ֶΛ࢖ͬͯ՝୊Λղܾ͢Δͱ͍͏ཁ੥΍໘ന͞
 ˠ ݺশͷ੝ਰΑΓ΋ͣͬͱ௕͘ଓͩ͘Ζ͏

  68. ੈք͔Β஌ࣝΛநग़͢ΔαΠΫϧ 68 ੈք σʔλ ஌ࣝ ॲཧ݁Ռ ஌ࣝΛͲ͏ੈքʹϑΟʔυόοΫ͢Δ͔ʁ ੈքʹͲ͏͋ͬͯ΄͍͔͠ͱ͍͏Ձ஋൑அ ԿΛɺԿͷͨΊʹͲ͏؍ଌ͢Δ͔ʁ ܭଌʹ஋͢Δ͔൱͔ͱ͍͏Ձ஋൑அ

  69. ର৅ͷੈքΛݶఆ͢Ε͹ࣗಈԽͷՄೳੑ͕͋Δ 69 ੈք σʔλ ஌ࣝ ॲཧ݁Ռ ໰୊ͷ੾Γग़͠ͱγεςϜԽ͸ɺιϑτ΢ΣΞΤϯδχΞϦϯάͷྖҬ ʢྫʣ঎඼΍هࣄͷਪનɺϚʔέςΟϯά Ձ஋൑அͷॏཁੑ͸࢒Γଓ͚Δ

  70. ֶੜͷօ͞Μ΁ͷϝοηʔδ 70 ֶͼଓ͚·͠ΐ͏ औΓ૊Έ·͠ΐ͏ - ֶ෦ɾେֶӃͰͷݚڀʢ՝୊ղܾͷαΠΫϧʣ ɹେֶ͸ੈքϨϕϧͷઐ໳Ո͔Βֶ΂Δوॏͳ৔ॴ - ਺ֶɾ޻ֶͷઐ໳஌ࣝ -

    ϓϩάϥϛϯά - ޠֶ - ٕज़ྙཧɺ๏੍౓ɺྺ࢙ɺ… ৬໊΍τϐοΫͷྲྀߦʹͱΒΘΕ͗ͣ͢ɺ ઐ໳஌ࣝͰ՝୊ղܾͰ͖ΔਓΛͥͻ໨ࢦ͍ͯͩ͘͠͞