Slide 1

Slide 1 text

͸͡ΊͯͷਓͷͨΊͷػցֶशೖ໳ ଜాݡଠ ΫοΫύουαϚʔΠϯλʔϯγοϓ

Slide 2

Slide 2 text

୭ʁ wଜాݡଠ ๺ւಓେֶത࢜ ৘ใՊֶ ‣ ॴଐ ‣ ձһࣄۀ෦αʔϏε։ൃάϧʔϓ ‣ ݚڀ։ൃνʔϜ ‣ 3VCZDPNNJUUFS ‣ ઐ໳෼໺ ‣ ෳࡶωοτϫʔΫɺػցֶश

Slide 3

Slide 3 text

ߨٛͷྲྀΕ ػցֶशͱ͸ ػցֶशͷ֓ཁ ڭࢣ͋ΓֶशͷྲྀΕ ·ͱΊ

Slide 4

Slide 4 text

ػցֶशͱ͸

Slide 5

Slide 5 text

Ṗͷ ٕज़ cat dog bicycle house meal https://www.flickr.com/photos/freefoto/4910780215 https://www.flickr.com/photos/onasill/16394567531 https://www.flickr.com/photos/muraken/19133489198 https://www.flickr.com/photos/muraken/19579394778 https://www.flickr.com/photos/muraken/15697878907 ػցֶशͷΠϝʔδ

Slide 6

Slide 6 text

ػցֶशͰͰ͖Δ͜ͱ wϨγϐ΁ͷࣗಈϥϕϧ෇͚ wҟͳΔϨγϐؒͷؔ࿈౓ͷਪఆ wϢʔβͷߦಈʹ߹Θͤͨίϯςϯπ഑৴ wࣸਅͷإೝࣝͱਓͷ൑ఆ wεύϜϝʔϧ൑ఆ w঎඼ͷਪન

Slide 7

Slide 7 text

ػցֶशͰͰ͖Δ͜ͱ w͞·͟·ͳϩά͔ΒͷσʔλϚΠχϯά ‣ ҩྍ਍அه࿥ ‣ 8FCͷΫϦοΫϩά ‣ αʔό΁ͷΞΫηεϩά ‣ FUD

Slide 8

Slide 8 text

ػցֶशͰͰ͖Δ͜ͱ wखॻ͖͢Δ͜ͱ͕ෆՄೳͳϓϩάϥϜ ‣ ϔϦ΍ࣗಈंͷࣗಈૢॎ ‣ खॻ͖จࣈೝࣝ ‣ ࣗવݴޠॲཧ ‣ ίϯϐϡʔλϏδϣϯ

Slide 9

Slide 9 text

ػցֶशͬͯʁ wਓ͕ؒࣗવʹߦ͍ͬͯΔֶशೳྗͱಉ༷ͷػೳΛ ίϯϐϡʔλͰ࣮ݱ͠Α͏ͱ͢Δٕज़ɾख๏
 Wikipedia w໌ࣔతʹϓϩάϥϜ͠ͳͯ͘΋ֶश͢ΔೳྗΛί ϯϐϡʔλʹ༩͑Δݚڀ෼໺
 ΞʔαʔɾαϛϡΤϧ, 1958

Slide 10

Slide 10 text

ػցֶशͬͯʁ wίϯϐϡʔλϓϩάϥϜ͕ɺ͋ΔछͷλεΫTͱ ධՁई౓Pʹ͓͍ͯܦݧE͔Βֶश͢Δͱ͸ɺλ εΫTʹ͓͚ΔͦͷϓϩάϥϜͷੑೳΛPʹΑͬ ͯධՁͨ͠ࡍʹܦݧEʹΑͬͯੑೳ͕վળ͞Εͯ ͍Δ৔߹Ͱ͋Δ
 τϜɾϛονΣϧ, 1998

Slide 11

Slide 11 text

ྫϨγϐ΁ͷࣗಈϥϕϧ෇͚ wϨγϐʹ෇͘ϥϕϧΛ༧ଌ͢Δ wϨγϐʹର͠Δద੾ͳϥϕϧΛਓ͕બͿ wϥϕϧ෇͚ͷ༧ଌ݁Ռ͕ਖ਼͔ͬͨ͠Ϩγϐͷ݅਺ λεΫ T ܦݧ E ධՁई౓ P

Slide 12

Slide 12 text

τϚτ 1 φε 0 ਫࡊ 0 ιό 1 ྫྷ΍͠ 1 : : Ϩγϐ ಛ௃ϕΫτϧ ෼ྨث : : 0.13 0.89 0.62 : ෼ྨ֬཰ϕΫτϧ ໙ྉཧ ओ৯ ೑ྉཧ : ෼ྨ݁Ռ Ϩγϐ΁ͷࣗಈϥϕϧ෇͚ͷྲྀΕ લॲཧ ޙॲཧ

Slide 13

Slide 13 text

ҟͳΔϨγϐؒͷؔ࿈౓ͷਪఆ Ϩγϐ A Ϩγϐ B Ϩγϐ C Ϩγϐͷಛ௃ϕΫτϧ ಛ௃ϕΫτϧؒͷ ྨࣅ౓ΛଌΔ ಛ௃ 1 ಛ௃ 2 ಛ௃ 3 ಛ௃ 4

Slide 14

Slide 14 text

Ϣʔβͷߦಈʹ߹Θͤͨίϯςϯπ഑৴ “ϙτϑ” Λ
 ݕࡧͨ͠ਓͷը໘ “ೲ౾” Λ
 ݕࡧͨ͠ਓͷը໘ Ϩγϐݕࡧ݁ՌͷதʹɺϨγϐͰ͸ͳ͍͕Ϩγϐ୳͠Λखॿ͚Ͱ͖Δ৘ใΛࠞͥͯ͋͛ͯɺϨγϐܾΊΛ ద੾ʹΞγετ͍ͨ͠ɻͲΜͳίϯςϯπΛࠞͥΔͱϢʔβ͸ຬ଍͢ΔΜͩΖ͏ʁ ໰୊: (޿ࠂ) “͓ʹ͗Βͣ” Λ
 ݕࡧͨ͠ਓͷը໘ (޿ࠂ)

Slide 15

Slide 15 text

Ṗͷٕज़͸Ͳ͏࣮ݱ ͞ΕͯΔΜͩΖ͏ʁ

Slide 16

Slide 16 text

ػցֶशͷ֓ཁ μ΢ϯϩʔυ͸ॱௐͰ͔͢ʁ

Slide 17

Slide 17 text

ػցֶशͷ࿮૊Έ wڭࢣ͋Γֶशsupervised le rning wڭࢣͳֶ͠शunsupervised le rning w൒ڭࢣ͋Γֶशsemi-supervised le rning wڧԽֶशreinforcement le rning

Slide 18

Slide 18 text

͖ͬ͞ྫΛ࿮૊Έʹ͋ͯ͸ΊΔ wࣸਅ΍ϨγϐͳͲ΁ͷࣗಈϥϕϧ෇͚ ‎ڭࢣ͋Γֶश wҟͳΔϨγϐؒͷؔ࿈౓ͷࣗಈਪఆ ‎ڭࢣͳֶ͠श wϢʔβͷߦಈʹ߹Θͤͨίϯςϯπ഑৴ ‎ڧԽֶश

Slide 19

Slide 19 text

ڭࢣ͋Γֶशͱ͸ wೖྗσʔλʹରͯ͠ग़ྗ͢΂͖ਖ਼ղσʔλ ڭࢣ σʔλ ͕༩͑ΒΕΔ ‣ ڭࢣσʔλϥϕϧ΍਺஋ͳͲ wਖ਼ղ͕෼͔Βͳ͍ೖྗσʔλʹରͯ͠ɺରԠ͢Δ ϥϕϧ΍਺஋Λ༧ଌ͢Δؔ਺΍نଇΛߏங͢Δ Ϟσϧ

Slide 20

Slide 20 text

ڭࢣ͋ΓֶशΛ࢖͏λεΫͷछྨ wճؼregression ‣ ࿈ଓ஋ͷग़ྗΛ༧ଌ͢ΔճؼϞσϧʢؔ਺ʣΛߏங wΫϥε෼ྨclassification ‣ ϥϕϧͷग़ྗΛ༧ଌ͢Δ෼ྨϞσϧΛߏங

Slide 21

Slide 21 text

ྫɿճؼϞσϧ

Slide 22

Slide 22 text

௚ઢʹΑΔճؼ 2࣍ۂઢʹΑΔճؼ ڭࢣσʔλ ༧ଌ݁Ռ

Slide 23

Slide 23 text

ྫɿ෼ྨϞσϧ BMI ˔ ˔ ˔ ˔ ˔ ˔ ʷ ʷ ʷ ʷ ʷ ʷ ʷ ˔ ˔ʜ݈߁ମ ʷʜ৺ଁප ෼ྨڥք (ՍۭͷσʔλͰ͢)

Slide 24

Slide 24 text

ڭࢣͳֶ͠शͱ͸ wೖྗσʔλʹରͯ͠ڭࢣσʔλ͸༩͑ΒΕͳ͍ wσʔλͷ෼෍ͳͲΛཔΓʹɺຊ࣭తͳߏ଄΍ύλʔ ϯΛநग़͢Δ

Slide 25

Slide 25 text

ڭࢣͳֶ͠शͱ͸ x1 x2 ˔ Ϩγϐ1 ˔ Ϩγϐ2 Ϩγϐ3 ˔ ˔ Ϩγϐ5 Ϩγϐ8 ˔ ˔ Ϩγϐ4 ˔ Ϩγϐ7 ˔ Ϩγϐ6 Ϩγϐ9 ˔ ˔ Ϩγϐ10 x1 x2 ˒ Ϩγϐ1 ˛ Ϩγϐ2 Ϩγϐ3 ˒ ˒ Ϩγϐ5 Ϩγϐ8 ˒ ˛ Ϩγϐ4 ˛ Ϩγϐ7 ˛ Ϩγϐ6 Ϩγϐ9 ˛ ˒ Ϩγϐ10 ڭࢣ͋Γֶश ڭࢣͳֶ͠श

Slide 26

Slide 26 text

ڭࢣͳֶ͠शͷछྨ wΫϥελϦϯά w࣍ݩ࡟ݮ wසग़ύλʔϯϚΠχϯά

Slide 27

Slide 27 text

ػցֶशͷྲྀΕ ܇࿅༻σʔλΛूΊΔ ΫϨϯδϯάͳͲͷલॲཧΛ͢Δ ಛ௃ʢૉੑʣͷઃܭΛ͢Δ ‣ χϡʔϥϧωοτϫʔΫͷ৔߹͸ӅΕ૚ͷઃܭΛ͢Δ ϞσϧΛֶश͠ɺݕূ͢Δ ‣ ݁Ռ͕ྑ͘ͳ͍৔߹͸PSʹ໭ͬͯ΍Γ௚͠ ӡ༻͢Δ

Slide 28

Slide 28 text

2. લॲཧ 3. ಛ௃ઃܭ ਤʹ͢ΔͱϦʔϯελʔτΞοϓΈ͍ͨͩͶ 4. Ϟσϧֶश 5. ݕূ 6. ӡ༻ 1. ܇࿅༻
 σʔλ

Slide 29

Slide 29 text

ػցֶशͷྲྀΕ ܇࿅༻σʔλΛूΊΔ ΫϨϯδϯάͳͲͷલॲཧΛ͢Δ ಛ௃ʢૉੑʣͷઃܭΛ͢Δ ‣ χϡʔϥϧωοτϫʔΫͷ৔߹͸ӅΕ૚ͷઃܭΛ͢Δ ϞσϧΛֶश͠ɺݕূ͢Δ ‣ ݁Ռ͕ྑ͘ͳ͍৔߹͸PSʹ໭ͬͯ΍Γ௚͠ ӡ༻͢Δ ໰୊ͷੑ࣭ʹ େ͖͘ґଘ͢Δ ໰୊ͷੑ࣭ • λεΫͷछྨ • ֶशσʔλͷྔ • ֶशσʔλͷ౷ܭతੑ࣭ • ͳͲ ໰୊ͷੑ࣭ʹґଘ ͢Δ෦෼͕͋Δ

Slide 30

Slide 30 text

ػցֶशͷྲྀΕ ܇࿅༻σʔλΛूΊΔ ΫϨϯδϯάͳͲͷલॲཧΛ͢Δ ಛ௃ʢૉੑʣͷઃܭΛ͢Δ ‣ χϡʔϥϧωοτϫʔΫͷ৔߹͸ӅΕ૚ͷઃܭΛ͢Δ ϞσϧΛֶश͠ɺݕূ͢Δ ‣ ݁Ռ͕ྑ͘ͳ͍৔߹͸PSʹ໭ͬͯ΍Γ௚͠ ӡ༻͢Δ ໰୊ͷੑ࣭ʹ͸
 ґଘ͠ͳ͍ ໰୊ͷੑ࣭ • λεΫͷछྨ • ֶशσʔλͷྔ • ֶशσʔλͷ౷ܭతੑ࣭ • ͳͲ

Slide 31

Slide 31 text

ڭࢣ͋ΓֶशͷྲྀΕ μ΢ϯϩʔυ͸ऴΓ·͔ͨ͠ʁ

Slide 32

Slide 32 text

τϚτ 1 φε 0 ਫࡊ 0 ιό 1 ྫྷ΍͠ 1 : : Ϩγϐ ಛ௃ϕΫτϧ ෼ྨث : : 0.13 0.89 0.62 : ෼ྨ֬཰ϕΫτϧ ໙ྉཧ ओ৯ ೑ྉཧ : ڭࢣσʔλ Ϩγϐ΁ͷࣗಈϥϕϧ෇͚༻෼ྨثͷֶश 0 1 1 : ޡࠩ ෼ྨύϥϝʔλͷमਖ਼

Slide 33

Slide 33 text

෼ྨثͷ࠷దԽޯ഑߱Լ๏ ݱࡏͷग़ྗ E(y) ޡࠩ y ग़ྗ ݱࡏͷޡࠩ : ޡࠩΛগ͠ݮগͤ͞ΔͨΊʹ
 ඞཁͳग़ྗͷඍখมԽྔ y ग़ྗΛมԽͤ͞ΔͨΊʹඞཁͳ
 ෼ྨύϥϝʔλͷमਖ਼ྔ y = f ( x ; ⇥) ͷͱ͖ y ⇥ : ෼ྨύϥϝʔλ ⇥ ⇥ = @E @⇥ = @E @y @f @⇥

Slide 34

Slide 34 text

൚Խೳྗͱաֶश w൚Խೳྗgener liz tion ‣ ܇࿅Ͱ࢖༻͍ͯ͠ͳ͍ະ஌ͷσʔλʹରͯ͠΋ޡΓ͕ খ͍͞༧ଌ͕Մೳͳ͜ͱ wաֶशʢաద߹ʣoverfitting ‣ ܇࿅Ͱ࢖༻ͨ͠σʔλʹద߹͗ͯ͢͠͠·͍ɺະ஌ͷ σʔλʹର͢Δྑ͍༧ଌ͕Ͱ͖ͳ͍͜ͱ

Slide 35

Slide 35 text

ྫɿճؼϞσϧ

Slide 36

Slide 36 text

௚ઢʹΑΔճؼ 2࣍ۂઢʹΑΔճؼ

Slide 37

Slide 37 text

2࣍ۂઢʹΑΔճؼ ߴ࣍ۂઢʹΑΔճؼ աֶशʢաద߹ʣ
 overfitting

Slide 38

Slide 38 text

ֶशͨ͠Ϟσϧͷݕূ wֶशͨ͠Ϟσϧͷ൚ԽೳྗΛ֬ೝ͢Δ wະ஌ͷೖྗΛਖ਼͘͠༧ଌͰ͖Δׂ߹ΛٻΊΔ

Slide 39

Slide 39 text

ڭࢣ͋Γֶशʹ͓͚Δݕূ wճؼͷ৔߹ wΫϥε෼ྨͷ৔߹

Slide 40

Slide 40 text

ճؼͷ৔߹ͷݕূ wਅ஋ͱ༧ଌ஋ͷࠩͷೋ৐࿨Λ࢖͏

Slide 41

Slide 41 text

Ϋϥε෼ྨͷ৔߹ͷݕূ wਖ਼͘͠෼ྨ͞Εׂͨ߹ͱޡͬͯ෼ྨ͞Εׂͨ߹Λ ར༻ͯ͠൚ԽೳྗΛݟੵΔ

Slide 42

Slide 42 text

ࠞ߹ߦྻconfusion m tri- ཅੑ ӄੑ ཅ
 ੑ ਅཅੑ 5SVF1PTJUJWF ِӄੑ 'BMTF/FHBUJWF ӄ
 ੑ ِཅੑ 'BMTF1PTJUJWF ਅӄੑ 5SVF/FHBUJWF ༧ଌͷ݁Ռ ਖ਼ղσʔλ ntp nfp nfn ntn ਖ਼ղ཰ accuracy ਅཅੑ཰ true positive rate ࠶ݱ཰ recall ਫ਼౓ʢద߹཰ʣ precision ntp ntp + nfn ntp + ntn ntp + nfp + ntn + nfn ntp ntp + nfp F-score 2 1 precision + 1 recall = n tp n tp + nfp+ nfn 2

Slide 43

Slide 43 text

ࠞ߹ߦྻconfusion m tri- ཅੑ ӄੑ ཅ
 ੑ ਅཅੑ 5SVF1PTJUJWF ِӄੑ 'BMTF/FHBUJWF ӄ
 ੑ ِཅੑ 'BMTF1PTJUJWF ਅӄੑ 5SVF/FHBUJWF ༧ଌͷ݁Ռ ਖ਼ղσʔλ ntp nfp nfn ntn ਖ਼ղ཰ accuracy ਅཅੑ཰ true positive rate ࠶ݱ཰ recall ਫ਼౓ʢద߹཰ʣ precision ntp ntp + nfn ntp + ntn ntp + nfp + ntn + nfn ntp ntp + nfp F-score 2 1 precision + 1 recall = n tp n tp + nfp+ nfn 2

Slide 44

Slide 44 text

ࠞ߹ߦྻconfusion m tri- ཅੑ ӄੑ ཅ
 ੑ ਅཅੑ 5SVF1PTJUJWF ِӄੑ 'BMTF/FHBUJWF ӄ
 ੑ ِཅੑ 'BMTF1PTJUJWF ਅӄੑ 5SVF/FHBUJWF ༧ଌͷ݁Ռ ਖ਼ղσʔλ ntp nfp nfn ntn ਖ਼ղ཰ accuracy ਅཅੑ཰ true positive rate ࠶ݱ཰ recall ਫ਼౓ʢద߹཰ʣ precision ntp ntp + nfn ntp + ntn ntp + nfp + ntn + nfn ntp ntp + nfp F-score 2 1 precision + 1 recall = n tp n tp + nfp+ nfn 2

Slide 45

Slide 45 text

ࠞ߹ߦྻconfusion m tri- ཅੑ ӄੑ ཅ
 ੑ ਅཅੑ 5SVF1PTJUJWF ِӄੑ 'BMTF/FHBUJWF ӄ
 ੑ ِཅੑ 'BMTF1PTJUJWF ਅӄੑ 5SVF/FHBUJWF ༧ଌͷ݁Ռ ਖ਼ղσʔλ ntp nfp nfn ntn ਖ਼ղ཰ accuracy ਅཅੑ཰ true positive rate ࠶ݱ཰ recall ਫ਼౓ʢద߹཰ʣ precision ntp ntp + nfn ntp + ntn ntp + nfp + ntn + nfn ntp ntp + nfp F-score 2 1 precision + 1 recall = n tp n tp + nfp+ nfn 2

Slide 46

Slide 46 text

Precision Recall F-score = 2 1 precision + 1 recall = n tp n tp + nfp+ nfn 2 Precision ͱ Recall ͷௐ࿨ฏۉ ͲͪΒ͔͕௿͍ͱ F-score ͸௿͍

Slide 47

Slide 47 text

ࠞ߹ߦྻconfusion m tri- ཅੑ ӄੑ ཅ
 ੑ ਅཅੑ 5SVF1PTJUJWF ِӄੑ 'BMTF/FHBUJWF ӄ
 ੑ ِཅੑ 'BMTF1PTJUJWF ਅӄੑ 5SVF/FHBUJWF ༧ଌͷ݁Ռ ਖ਼ղσʔλ ntp nfp nfn ntn ਖ਼ղ཰ accuracy ਅཅੑ཰ true positive rate ࠶ݱ཰ recall ਫ਼౓ʢద߹཰ʣ precision ntp ntp + nfn ntp + ntn ntp + nfp + ntn + nfn ntp ntp + nfp F-score 2 1 precision + 1 recall = n tp n tp + nfp+ nfn 2 F-score ͕ߴ͚Ε͹ྑ͍ͷʁ → ໰୊ʹґଘ͢Δ

Slide 48

Slide 48 text

ྫɿҩྍσʔλͷ෼ྨ BMI ˔ ˔ ˔ ˔ ˔ ˔ ʷ ʷ ʷ ʷ ʷ ʷ ʷ ˔ ˔ʜ݈߁ମ ӄੑ ʷʜ৺ଁප ཅੑ (ՍۭͷσʔλͰ͢) ِӄੑ ِཅੑ ҩྍσʔλͷ෼ྨͰ͸ِӄੑΛ0݅ʹ͍ͨ͠ → ࠶ݱ཰Λ100%ʹ͢Δ͜ͱ͕ॏཁ

Slide 49

Slide 49 text

ྫɿҩྍσʔλͷ෼ྨ BMI ˔ ˔ ˔ ˔ ˔ ˔ ʷ ʷ ʷ ʷ ʷ ʷ ʷ ˔ ˔ʜ݈߁ମ ӄੑ ʷʜ৺ଁප ཅੑ (ՍۭͷσʔλͰ͢) ِӄੑ ِཅੑ ҩྍσʔλͷ෼ྨͰ͸ِӄੑΛ0݅ʹ͍ͨ͠ → ࠶ݱ཰Λ100%ʹ͢Δ͜ͱ͕ॏཁ

Slide 50

Slide 50 text

ଟΫϥε෼ྨͷ৔߹ wΫϥεຖʹࠞ߹ߦྻΛ࡞Γ౷߹͢Δ

Slide 51

Slide 51 text

Ϋϥεͷ৔߹ ཅੑ ӄੑ ཅੑ ӄੑ ༧ଌͷ݁Ռ ਖ਼ղσʔλ ཅੑ ӄੑ ཅੑ ӄੑ ਖ਼ղσʔλ ཅੑ ӄੑ ཅੑ ӄੑ ਖ਼ղσʔλ n(1) tp n(2) tp n(3) tp n(1) tn n(2) tn n(3) tn n(3) fp n(2) fp n(1) fp n(1) fn n(2) fn n(3) fn Ϋϥε1 Ϋϥε2 Ϋϥε3 ֤Ϋϥεʹଐ͢Δ ֤Ϋϥεʹଐ͞͵ ༧ଌͷ݁Ռ ਖ਼ղσʔλ ֤Ϋϥεʹ
 ଐ͢Δ ֤Ϋϥεʹ
 ଐ͞͵ n(1) fn + n(2) fn + n(3) fn n(1) tp + n(2) tp + n(3) tp n(1) fp + n(2) fp + n(3) fp n(1) tn + n(2) tn + n(3) tn

Slide 52

Slide 52 text

ަࠩݕূcross v lid tion w܇࿅σʔλΛֶश༻ͱݕূ༻ʹ෼ׂ͢Δύλʔϯ Λมߋ͠ɺෳ਺ͷݕূ݁ՌͷฏۉΛͱΔ͜ͱͰɺ ൚ԽੑೳΛ෼ׂύλʔϯʹґଘ͠ͳ͍ͰଌΔ ‣ K-෼ׂަࠩݕূ ‣ LOOCV (Leave-one-out ަࠩݕূ)

Slide 53

Slide 53 text

,෼ׂަࠩݕূ 1 2 3 4 … K N ݸͷֶशσʔλΛ K ϒϩοΫʹ෼ׂ ݕূʹ࢖͏ϒϩοΫΛ ॱ൪ʹ੾Γସ͑ͯ K ύλʔϯͷݕূΛߦ͏ ֶश༻ϒϩοΫ ݕূ༻ϒϩοΫ (K=N ͷͱ͖ LOOCV ʹͳΔ)

Slide 54

Slide 54 text

,ͷબͿͱ͖ʹؾʹ͢Δ͜ͱ ֶश༻σʔλ ݕূ༻
 σʔλ K=2 K=N N ݕূύλʔϯ਺ 2 ≒ܭࢉίετ N(K 1) K K = 2 K = N N 2

Slide 55

Slide 55 text

·ͱΊ

Slide 56

Slide 56 text

·ͱΊ wػցֶशͱ͸ɺίϯϐϡʔλϓϩάϥϜ͕ܦݧʹ ΑͬͯλεΫͷղ͖ํΛֶΜͰ͍͘࢓૊Έͷ͜ͱ wػցֶशΛ༻͍ͨγεςϜ͸ɺαʔϏε։ൃͱಉ ͡Α͏ͳαΠΫϧͰ։ൃɾӡ༻͞ΕΔ wϞσϧͷ൚Խೳྗ͕ॏཁͰ͋ΔͨΊɺաֶशͯ͠ ͍ͳ͍ࣄΛݕূ͢Δඞཁ͕͋Δ