effectsパッケージを用いた一般化線形モデルの可視化 /NagoyaR15

by Yu Tamura

Slide 1

Slide 1 text

effectsύοέʔδΛ༻͍ͨ ҰൠԽઢܗϞσϧͷՄࢹԽ 2016.03.27 Nagoya.R #15 ໊ݹ԰େֶେֶӃࠃࡍ։ൃݚڀՊ D2 ాଜ༞ 1

Slide 2

Slide 2 text

֓ཁ • ର৅ • ઢܗϞσϧɼҰൠԽઢܗϞσϧΛ࢖ͬͨ෼ੳΛ ͢Δ͜ͱ͕͋Δํ • ಺༰ • ্هͷΑ͏ͳ෼ੳͷ݁ՌͷՄࢹԽΛศརʹ΍ͬ ͯ͘ΕΔeffectsύοέʔδͷ঺հ 2

Slide 3

Slide 3 text

ීஈ͸͠ͳ͍ࣗݾ঺հ • ໊લ • ాଜ༞ • ॴଐ • ໊ݹ԰େֶେֶӃࠃࡍ։ൃݚڀՊത࢜ޙظ՝ఔ • ݚڀ෼໺ • ୈೋݴޠशಘɼ৺ཧݴޠֶɼจ๏ࢦಋ • Rྺ • ҰੜΤϯυϢʔβʔʢ3೥͘Β͍ʣ 3

Slide 4

Slide 4 text

͓அΓ • ࠓճ͸ɼͱ͘ʹࣗ෼ʹؔ܎ͷ͋ΔݚڀͷσʔλΛ ࢖͍·ͤΜʢ४උෆ଍ŗŖŕʣ • effectsύοέʔδʹೖ͍ͬͯΔαϯϓϧͷσʔλ ͷத͔Β2ͭΛ࢖͍·͢ 4

Slide 5

Slide 5 text

ҰݴͰݴ͏ͱ • ෼ੳΛͨ͠ΒͱΓ͋͑ͣ • plot(alleffects(model)) • ͱ΍ͬͯΈ·͠ΐ͏ 5

Slide 6

Slide 6 text

αϯϓϧσʔλᶃ • Cowles • Cowles, M., & Davis, C. (1987). The subject matter of psychology: Volunteers. British Journal of Social Psychology, 26, 97–102. • 4ม਺ • neuroticismʢ৘ॹෆ҆ఆੑʣ • extraversionʢ֎޲ੑʣ • sexʢੑผʣ • volunteerʢࠓޙͷௐࠪʹࣗओతʹࢀՃ͢Δ͔ʣ • Yes or Noͷ2୒ 6 ΞΠθϯΫੑ֨ ݕࠪͷείΞ

Slide 7

Slide 7 text

த਎ΛݟͯΈΔ 7

Slide 8

Slide 8 text

෼ੳͯ͠ΈΔ • ੑผɼ֎޲ੑɼ৘ॹෆ҆ఆੑ͕ɼௐࠪ΁ͷࣗओ తࢀՃʹͲͷΑ͏ʹӨڹ͢Δͷ͔Λௐ΂͍ͨ 8 ैଐม਺ ಠཱม਺

Slide 9

Slide 9 text

෼ੳͯ͠ΈΔ • 2஋ม਺Λैଐม਺ͱͨ͠ϩδεςΟοΫճؼ෼ ੳ • ͱΓ͋͑ͣɼ3ͭͷಠཱม਺͢΂ͯͷओޮՌͱަ ޓ࡞༻ΛϞσϧʹ૊ΈࠐΉ 9

Slide 10

Slide 10 text

10 ΞελϦεΫͰͭͳ͙ͱओޮՌˍަޓ࡞༻ binomial͸ೋ߲෼෍ͷ͜ͱ ŗƀŕ

Slide 11

Slide 11 text

෼ੳͯ͠ΈΔ • 3࣍ͷަޓ࡞༻͸ͳ͛͞ͳͷͰɼAICͰϞσϧબ ୒ • glmؔ਺Λ࢖ͬͨϞσϦϯάͷ৔߹͸stepAIC͕ ࢖͑ΔͷͰศར 11

Slide 12

Slide 12 text

Slide 13

Slide 13 text

෼ੳͯ͠ΈΔ • Θ͔ͬͨ͜ͱ • ੑผͷओޮՌ͋Γ • ৘ॹෆ҆ఆੑͱ֎޲ੑͷަޓ࡞༻͋Γ ※ͪͳΈʹࢀՃऀIDΛϥϯμϜ੾ยʹͨ͠GLMM΋΍ͬͯΈ·͕ͨ͠ਪఆ ͕͏·͘Ͱ͖·ͤΜͰͨ͠ 13

Slide 14

Slide 14 text

ͱݴΘΕͯ΋ ϐϯͱ͜ͳ͍ͷͰ 14

Slide 15

Slide 15 text

ਤࣔͯ͠ΈΔ >eff.cowles <-allEffects(model2, xlevels =list(extraversion = seq(0,24,6)),given.values=c(sexmale = 0.5)) 15 ͜ͷதʹGLMͷग़ ྗ͕ೖ͍ͬͯΔ xlevelsͰx࣠ͷ໨੝Γ Λࢦఆ ۠੾Γͷࢦఆɻʮ0͔Β 24·ͰΛ6ͣͭʯҙຯ ͜͏͢Δ͜ͱͰɼϞσϧ ͷத͔Βਤࣔʹඞཁͱͳ ΔΦϒδΣΫτΛ࡞Δ

Slide 16

Slide 16 text

ਤࣔͯ͠ΈΔ • த਎͕Ͳ͏ͳ͍ͬͯΔͷ͔ݟͯΈΔͱ 16

Slide 17

Slide 17 text

ਤࣔͯ͠ΈΔ • ͋ͱ͸ඳ͚ͩ͘ >plot(eff.cowles, ylab = “Prob(Volunteer)”) 17

Slide 18

Slide 18 text

Slide 19

Slide 19 text

19 extraversion = 0 extraversion = 6 extraversion = 12 extraversion = 18 extraversion = 24

Slide 20

Slide 20 text

ิ଍ • allEffects()͸ɼϞσϧͷதͷશͯͷ߲ΛऔΓग़͢ ৔߹ʹ࢖͏ • ϞσϧͷதͷҰ෦ͷΈͰྑ͍৔߹͸ɼeffect·ͨ ͸EffectΛ࢖ͬͯҎԼͷΑ͏ʹͰ͖Δ >eff.ne <-effect(“neuroticism*extraversion”,model2) >Eff.ne <-Effect(c(“neuroticism”, “extraversion”), model2) 20

Slide 21

Slide 21 text

ิ଍ • multiline = TRUEͷࢦఆΛ͢Ε͹̍ͭͷਤʹ >plot(eff.cowles,”sex”,ylab=“Prob(Volunteer)”) 21

Slide 22

Slide 22 text

ิ଍ • ਤΛݸผʹඳ͖͍ͨ৔߹͸߲Λࢦఆ͢Ε͹OK >plot(eff.cowles,”neuroticism:extraversion”,m ultiline = T,ylab=“Prob(Volunteer)”) • Ұ౓ม਺ʹೖΕͣʹplot಺ͰೖΕࢠͷܗʹ͢Ε͹ ಉ͡ਤʹͳΔ >plot(effect(“neuroticism:extraversion”,model 2,xlevels = seq(0,24,6))),multiline =T) 22

Slide 23

Slide 23 text

Slide 24

Slide 24 text

αϯϓϧσʔλᶄ • BEPSʢBritish Election Panel Study) • effectsύοέʔδʹೖ͍ͬͯΔαϯϓϧσʔλ • 1997-2001೥ͷબڍʹؔ͢ΔΠΪϦεͷௐࠪ σʔλ 24

Slide 25

Slide 25 text

• 10ม਺ • vote • Ͳͷ੓ౘʹ౤ථ͔ͨ͠ • Conservative, Labour, Liberal Democratͷ3ͭ • age • ೥ྸ • economi.cond.national • ݱࡏͷࠃͷܦࡁঢ়گʢ1-5Ͱධఆʣ • economic.cond.household • ݱࡏͷՈܭͷܦࡁঢ়گʢ1-5Ͱධఆʣ • Blair • ࿑ಇౘౘटBlairͷධՁʢ1-5Ͱධఆʣ • Hague • อकౘౘटHagueͷධՁʢ1-5Ͱධఆʣ • Kennedy • ࣗ༝ຽओౘౘटKennedyͷධՁʢ1-5Ͱධఆʣ • Europe • Ԥभջٙओٛʹର͢Δଶ౓ʢ1-11Ͱධఆʣ • ਺஋͕ߴ͍΄ͲԤभջٙओٛత • political.knowledge • ͦΕͧΕͷౘ͕Ԥभջٙओٛʹରͯ͠Ͳ͏͍ͬͨϙδγϣϯΛऔ͍ͬͯΔ͔ʹؔ͢Δ஌ࣝʢ0-3 Ͱධఆʣ • gender • male͔female͔ͷ2ਫ४ 25

Slide 26

Slide 26 text

த਎ΛݟͯΈΔ 26 ͪͬ͞

Slide 27

Slide 27 text

෼ੳͯ͠ΈΔ • ͞·͟·ͳม਺͕౤ථߦಈʹͲͷΑ͏ͳӨڹΛ ༩͍͑ͯΔͷ͔ 27

Slide 28

Slide 28 text

Slide 29

Slide 29 text

෼ੳͯ͠ΈΔ • ΧςΰϦΧϧม਺ʢڞ࿨ౘɾࣗຽౘɾ࿑ಇౘʣ Λैଐม਺ʹͨ͠ଟ߲ϩδεςΟοΫճؼ • nnetύοέʔδͷmultinomؔ਺Λ࢖༻ 29

Slide 30

Slide 30 text

30 ͪͬ͞

Slide 31

Slide 31 text

ਤࣔͯ͠ΈΔ • ͱΓ͋͑ͣ̍ͭͣͭΈͯΈΔ >plot(effect("age",beps),ylab="vote(Probability ") 31 ࣗຽౘ ࿑ಇౘ อकౘ

Slide 32

Slide 32 text

ਤࣔͯ͠ΈΔ • ࠃͷܦࡁঢ়گͷධՁͱ౤ථ཰ >plot(effect(“economic.cond.national”,beps),yl ab="vote(Probability") 32 ࣗຽౘ ࿑ಇౘ อकౘ

Slide 33

Slide 33 text

ਤࣔͯ͠ΈΔ • Ոܭͷܦࡁঢ়گͱ౤ථ཰ >plot(effect(“economic.cond.household”,beps) ,ylab="vote(Probability") 33 ࣗຽౘ ࿑ಇౘ อकౘ

Slide 34

Slide 34 text

ਤࣔͯ͠ΈΔ • ౘटͷධՁ͸Ͳ͏͔ >plot(effect(“Blair”,beps),ylab=“vote(Probabilit y") >plot(effect(“Hague”,beps),ylab=“vote(Probabil ity") >plot(effect(“Kennedy”,beps),ylab="vote(Proba bility") 34

Slide 35

Slide 35 text

Slide 36

Slide 36 text

Slide 37

Slide 37 text

37 Balir(࿑ಇౘ) Hague(อकౘ) Kennedy(ࣗຽౘ) ͜Εʹؾ͍ͮͨํ͸Rݕఆ4ڃͰ͢ʢӕ

Slide 38

Slide 38 text

"OZXBZ

Slide 39

Slide 39 text

39 x࣠ political.knowledge௿ x࣠ political.knowledgeߴ ύωϧ Ԥभջٙओٛత౓߹͍௿ ύωϧ Ԥभջٙओٛత౓߹͍ߴ

Slide 40

Slide 40 text

ิ଍ᶄ • plot͢Δͱ͖ͷࡉ͔͍Ҿ਺ࢦఆ • band.colors :৴པ۠ؒͷόϯυͷ৭ʢσϑΥͰ փ৭ʣ • band.transparency:όϯυͷಁ͚ಁ͚۩߹ • ci.style: “bands”, “bars”, “lines”Ͱબ΂Δʢͬ͞ ͖·Ͱͷ΍ͭ͸”bands”Ͱ͢ʣ 40

Slide 41

Slide 41 text

41 band.colors="red",band.transparency=0.3,ci.style = "lines"

Slide 42

Slide 42 text

͓ΘΓʹ • ਤΛࡉ͔͍͘͡ΔͷͳΒ΍͸Γ௿ਫ४࡞ਤؔ਺ ͳͲΛγίγίͱۦ࢖ͨ͠ΓɼggplotΛ࢖ͬͯඳ ͘΄͏͕ྑ͍ͷ͔΋ʢݟͨ໨ʹͩ͜ΘΔํʣ • ͨͩ͠ɼeffΦϒδΣΫτΛ࡞Δ࡞ۀ͚ͩ΍ͬͯ ͓͍ͯɼͦͷޙʹggplotͰਤࣔͱ͍͏ύλʔϯ΋ Ͱ͖Δʢࠓճ͸ׂѪʣ 42

Slide 43

Slide 43 text

ࢀߟจݙ Cowles, M., & Davis, C. (1987). The subject matter of psychology: Volunteers. British Journal of Social Psychology, 26, 97–102. Fox, J. (2003) Effect displays in R for generalised linear models. Journal of Statistical Software, 8, 1–27, doi: 10.18637/jss.v008.i15 Fox, J., Weisverg, S., Friendly, M., & Hong, J (2016) effects [R package version 3.0-7] Retrieved from https://cran.r-project.org/ package=effects 43

Slide 44

Slide 44 text

effectsύοέʔδΛ༻͍ͨ ҰൠԽઢܗϞσϧͷՄࢹԽ contact info ాଜ ༞ ໊ݹ԰େֶେֶӃੜ [email protected] http://www.tamurayu.wordpress.com/ 44 ෼ੳͨ͠ΒͱΓ͋͑ͣ >plot(allEffects(model1)) ͯ͠ΈΔ