Share
Embed
数学カフェ 確率統計入門
2017/4/22(土)【第18回数学カフェ】【確率・統計・機械学習回】
https://connpass.com/event/54301/
「速習 確率・統計」
本講演では、主に確率、統計の初心者の方や、プログラマ、エンジニアの方でデータ分析に興味を持っている方が確率統計のエッセンスを数理的に理解できることを目標に、データの集約方法から、大数の法則や中心極限定理など、確率・統計で利用される非常に重要な数学の定理などを紹介します。
kenmatsu4
November 29, 2023
Transcript
228 slides
-
ࣗݾհ!LFONBUTV ɾ'BDFCPPLϖʔδ ɹIUUQTXXXGBDFCPPLDPNNBUTVLFOCPPL ɾ5XJUUFSΞΧϯτ ɹ!LFONBUTV ɾ2JJUBͰϒϩάΛॻ͍͍ͯ·͢ʢ౷ܭɺػցֶशɺ1ZUIPOʣ ɹɹɹIUUQRJJUBDPNLFONBUTV ɹɹɹʢDPOUSJCVUJPOΛ͑·ͨ͠ʂʣ ɾझຯ ɹɹɹόϯυͰϕʔεΛ͍ͨΓ͍ͯ͠·͢ɻ
ɹɹɹओʹ౦ೆΞδΞόοΫύοΫཱྀߦʹߦͬͨΓ͠·͢ ɹɹʢΧϯϘδΞɺϛϟϯϚʔɺόϯάϥσγϡɺ৽ᙜΠάϧ࣏ࣗ۠FUDʣ ཱྀߦͷࣸਅIUUQNBUTVLFOKJNEPDPN TwitterΞΠίϯ -
౷ܭֶ ΧςΰϦ λΠτϧ Ұൠ ʲ౷ܭֶʳॳΊͯͷʮඪ४ภࠩʯʢ౷ܭֶʹ࠳ં͠ͳ͍ͨΊʹʣ Ұൠ ౷ܭֶجૅͷجૅ Ұൠ ෳࡶոحͳਖ਼نͷࣜͷҙຯΛಡΈղ͘ Ұൠ
ʲ౷ܭֶʳਖ਼نͱΧΠೋͷؔΛՄࢹԽͯ͠ΈΔɻ Ұൠ ʲ౷ܭֶʳத৺ۃݶఆཧͷΠϝʔδΛάϥϑͰ௫Ή Ұൠ ʲ౷ܭֶʳͬͯԿʁΛάϥϑΟΧϧʹઆ໌ͯ͠ΈΔɻ Ұൠ ʲ౷ܭֶʳ30$ۂઢͱԿ͔ɺΞχϝʔγϣϯͰཧղ͢Δɻ Ұൠ ʲ౷ܭֶʳ22ϓϩοτͷΈΛΞχϝʔγϣϯͰཧղ͢Δɻ ճؼੳ 1ZUIPOΛͬͨճؼੳͷ֓೦ͷղઆͦͷ̍ ճؼੳ QZUIPOΛͬͨճؼੳͷ֓೦ͷղઆͦͷ̎ ճؼੳ 1ZUIPOΛͬͨճؼੳͷ֓೦ͷղઆ൪֎ฤ̍ ճؼੳ ʲ౷ܭֶʳҰൠԽઢܗࠞ߹Ϟσϧ (-.. Λཧղ͢ΔͨΊͷՄࢹԽɻ ճؼੳ ʲ౷ܭֶʳʲ3ʳҐճؼΛͬͯΈΔɻ ճؼੳ ʲ౷ܭֶʳ֎Ε͕͍͍͋ͬͯײ͡ͰճؼઢΛҾ͍ͯΈΔ .$.$ ʲ౷ܭֶʳϚϧίϑ࿈ϞϯςΧϧϩ๏ .$.$ ʹΑΔαϯϓϦϯάΛΞχϝʔγϣϯͰղઆͯ͠ΈΔɻ .$.$ ʲ౷ܭֶʳϋϛϧτχΞϯϞϯςΧϧϩ๏ΛΞχϝʔγϣϯͰՄࢹԽͯ͠ཧղ͢Δɻ .$.$ ༁1Z.$5VUPSJBMʢϚϧίϑ࿈ϞϯςΧϧϩ๏ϑϨʔϜϫʔΫʣͦͷ̍ .$.$ ༁1Z.$5VUPSJBMʢϚϧίϑ࿈ϞϯςΧϧϩ๏ϑϨʔϜϫʔΫʣͦͷ̎ʮ$BTFTUVEZ֬తϘϥ ςΟϦςΟϞσϧʯ .$.$ ʲ౷ܭֶʳTUBOͰϩδεςΟοΫճؼͷ࣮ߦΛׂͱৄ͘͠ղઆͯ͠ΈΔ X5JUBOJDEBUBTFU .$.$ ʲ౷ܭֶʳ.$.$αϯϓϦϯάͷ.VMUJQSPDFTTJOHԽ ओੳ ओੳ1$"Λ༻͍ͯखॻ͖ࣈΛੳ͢Δɻͦͷ̍ ओੳ ओੳ1$"Λ༻͍ͯखॻ͖ࣈΛੳ͢Δɻͦͷ̎ ࣌ܥྻղੳ ʲ౷ܭֶʳ<࣌ܥྻղੳ>"3."ϞσϧΛͻͨ͢Βϓϩοτͯ͠Λ͔ͭΉɻ ػցֶश ΧςΰϦ λΠτϧ Ұൠ ֬తޯ߱Լ๏ͱԿ͔ɺΛ1ZUIPOͰಈ͔ͯ͠ղઆ͢Δ Ұൠ ʲػցֶशʳ-1ϊϧϜͬͯͳΜ͚ͩͬʁ Ұൠ ʲػցֶशʳϞσϧධՁɾࢦඪʹ͍ͭͯͷ·ͱΊͱ࣮ߦ X5JUBOJDσʔληοτ Ұൠ ʲػցֶशʳ00# 0VU0G#BH ͱͦͷൺ खॻ͖ࣈ खॻ͖ࣈΛQZUIPOͰͯ͋ͦͿͦͷ̍ खॻ͖ࣈ खॻ͖ࣈΛQZUIPOͰͯ͋ͦͿͦͷ̎ʢࣝผ͢Δʣ खॻ͖ࣈ ʲػցֶशʳLOFBSFTUOFJHICPSNFUIPE L࠷ۙ๏ ΛࣗྗͰQZUIPOͰॻ͍ͯɺखॻ͖ࣈ ͷೝࣝΛ͢Δ %FFQ-FBSOJOH ʲػցֶशʳσΟʔϓϥʔχϯάϑϨʔϜϫʔΫ$IBJOFSΛࢼ͠ͳ͕Βղઆͯ͠ΈΔɻ %FFQ-FBSOJOH ʲσΟʔϓϥʔχϯάʳ$IBJOFSͰ"VUPFODPEFSΛࢼͯ݁͠ՌΛՄࢹԽͯ͠ΈΔɻ 4QBSL.-MJC ʲػցֶशʳJ1ZUIPO/PUFCPPLͰ4QBSLΛىಈͤͯ͞.-MJCΛࢼ͢ 4QBSL.-MJC ʲػցֶशʳ4QBSL.-MJCΛ1ZUIPOͰಈ͔ͯ͠Ϩίϝϯσʔγϣϯͯ͠ΈΔ 4QBSL.-MJC ʲػցֶशʳ:BIPP/FXTͷهࣄΛ.-MJCͷτϐοΫϞσϧ -%" ͰΫϥελϦϯά͢Δɻ ҟৗݕ ʲػցֶशʳʮҟৗݕͱมԽݕষʯωΠϚϯϐΞιϯͷิͷࣜͷߦؒΛຒΊͯΈΔ ҟৗݕ ʲػցֶशʳʮҟৗݕͱมԽݕʯ$IBQUFSͷਤΛ1ZUIPOͰඳ͍ͯΈΔɻ ҟৗݕ 3ͷີൺਪఆύοέʔδEFOTSBUJPΛ1ZUIPO͔Βར༻͢Δ εύʔε ʲ1Z4UBOʳ(SBQIJDBM-BTTPΛ4UBOͰͬͯΈΔ ·ͱΊ ػցֶशϓϩϑΣογϣφϧγϦʔζྠಡձεϥΠυ·ͱΊ ϓϩάϥϛϯά ΧςΰϦ λΠτϧ 5XJUUFS ελόͷ5XJUUFSσʔλΛQZUIPOͰେྔʹऔಘ͠ɺσʔλੳΛࢼΈΔͦͷ̍ 5XJUUFS ελόͷ5XJUUFSσʔλΛQZUIPOͰେྔʹऔಘ͠ɺσʔλੳΛࢼΈΔͦͷ̎ 5XJUUFS ελόͷ5XJUUFSσʔλΛQZUIPOͰେྔʹऔಘ͠ɺσʔλੳΛࢼΈΔͦͷ̏ 5XJUUFS ελό5XJUUFSσʔλҐஔใͷϏδϡΞϥΠθʔγϣϯͱੳ 5XJUUFS 5XJUUFS4USFBN"1*σʔλʹରͯ͠ॳาతͳײੳΛࢼΈΔɻ 5XJUUFS ੈքͰ࠷Өڹྗͷ͋Δਓͷٕज़ܥπΠολʔϢʔβʔใΛQZUIPOͰऔಘ͢Δɻ ศརா ࢲత1ZUIPOศརா ਵ࣌ߋ৽ άϥϑΟΫε .BDͰ1ZUIPO͔ΒΞχϝʔγϣϯ(*'Λੜ͢Δڥઃఆ άϥϑΟΫε ʲ1ZUIPOʳNBUQMPUMJCʹΑΔάϥϑඳը࣌ͷ$PMPSNBQͷΧελϚΠζ $ZUIPO +VQZUFS/PUFCPPLͰ$ZUIPOΛ͏<1ZUIPO> 8PSE$MPVE 8PSE$MPVEͰจষͷ୯ޠग़ݱසΛՄࢹԽ͢Δɻ<1ZUIPO> άϥϑ%# 1ZUIPOͰάϥϑσʔλϕʔε/FPKೖGPSϏΪφʔ .BD049͚ +VMJB +VQZUFSͰ+VMJBΛಈ͔ͯ͠ճؼੳΛͬͯΈΔɻ ֶɺͦͷଞ ΧςΰϦ λΠτϧ ֶ ʲֶʳʮੵʯͷҙຯΛάϥϑΟΧϧʹཧղ͢Δͱ৭ʑݟ͑ͯ͘Δͦͷ̍ ֶ ʲֶʳݻ༗ɾݻ༗ϕΫτϧͱԿ͔ΛՄࢹԽͯ͠ΈΔ ֶ ΠΣϯηϯ +FOTFO ͷෆࣜͷײతཧղ ֶ ϐβͰཧղ͢ΔͷׂΓࢉͷҙຯ ·ͱΊ ʲ2JJUB"1*ʳ<౷ܭֶwػցֶश>ࠓ·Ͱͷߘهࣄͷ·ͱΊͱੳͬͯΈͨɻ 2JJUBߘهࣄҰཡ -
ΧςΰϦ λΠτϧ 63- ౷ܭֶ جૅ͔ΒͷϕΠζ౷ܭֶྠಡձࢿྉୈষʮൺɾ૬ؔɾ৴པੑʯ IUUQTXXXTMJEFTIBSFOFUNBUTVLFOCPPL ౷ܭֶ ୈ̍̏ճֶΧϑΣʮૉʂʂʯೋ࣍ձ-5ࢿྉʮཚʂʂʯ IUUQTXXXTMJEFTIBSFOFUNBUTVLFOCPPLMUSFW ౷ܭֶ
ʮϕʔλͷಾʹഭΔʯୈճϓϩάϥϚͷͨΊͷֶษڧձ-5ࢿྉ IUUQTXXXTMJEFTIBSFOFUNBUTVLFOCPPLMU ౷ܭֶ ͰΘ͔Δ ൣғͷ ϕΠζ౷ܭֶ IUUQTXXXTMJEFTIBSFOFUNBUTVLFOCPPL ౷ܭֶ جૅ͔ΒͷϕΠζ౷ܭֶྠಡձࢿྉୈষϝτϩϙϦεɾϔΠεςΟϯάε๏ IUUQTXXXTMJEFTIBSFOFUNBUTVLFOCPPL ౷ܭֶ ʮશͯͷ֬ίΠϯ͚͛ʹ௨ͣʯ+BQBO3ൃදࢿྉ IUUQTXXXTMJEFTIBSFOFUNBUTVLFOCPPLKBQBOS ౷ܭֶ جૅ͔ΒͷϕΠζ౷ܭֶྠಡձࢿྉୈষ֬ʹؔ͢ΔϕΠζͷఆཧ IUUQTXXXTMJEFTIBSFOFUNBUTVLFOCPPL ౷ܭֶ ౷ܭֶͷجૅͷجૅ IUUQTXXXTMJEFTIBSFOFUNBUTVLFOCPPLTT ػցֶश /FVSBM/FUXPSLΛՄࢹԽͯ͠ΈΔʢֶΧϑΣձ-5ࢿྉʣ IUUQTXXXTMJEFTIBSFOFUNBUTVLFOCPPLOFVSBMOFUXPSLMU ػցֶश %4-5ࡇΓʮ"6$͕վળͨͬͯ͠Ͳ͏͍͏͜ͱͰ͔͢ʁʯ IUUQTXXXTMJEFTIBSFOFUNBUTVLFOCPPLETMUBVD ػցֶश ౷ܭతֶशͷجૅষલ IUUQTXXXTMJEFTIBSFOFUNBUTVLFOCPPL ػցֶश 3BOEPN'PSFTUʹΑΔྨ IUUQTXXXTMJEFTIBSFOFUNBUTVLFOCPPLSBOEPN GPSFTU ػցֶश ҟৗݕͱมԽݕୈষۙ๏ʹΑΔҟৗݕ IUUQTXXXTMJEFTIBSFOFUNBUTVLFOCPPL ػցֶश ʮਂֶशʯษڧձ-5ࢿྉ$IBJOFSͬͯΈͨ IUUQTXXXTMJEFTIBSFOFUNBUTVLFOCPPLMUDIBJOFS ػցֶश %FFQ-FBSOJOH$IBQ$POWPMVUJPOBM/FVSBM/FU IUUQTXXXTMJEFTIBSFOFUNBUTVLFOCPPLEFFQMFBSOJOHDIBQ DPOWPMVUJPOBMOFVSBMOFU ػցֶश ʮਂֶशʯୈ̒ষࠐΈχϡʔϥϧωοτ IUUQTXXXTMJEFTIBSFOFUNBUTVLFOCPPLTT ֶ ʮੵ͕ݟ͑Δͱ౷ܭֶݟ͑ΔʯୈճϓϩάϥϚͷͨΊͷֶษڧձൃදࢿྉ IUUQTXXXTMJEFTIBSFOFUNBUTVLFOCPPL ͦͷଞ ന͍ཅؾͳͭͱ1ZUIPOͰͨΘΉΕΔ IUUQTXXXTMJEFTIBSFOFUNBUTVLFOCPPLQZUIPO 4MJEFTIBSFߘهࣄҰཡ -
ຊൃදͷ༰ େͷ๏ଇ ਖ਼ن த৺ۃݶఆཧ ύϥϝʔλʔͷਪఆͱ ࠷๏ ۠ؒਪఆ ౷ܭతݕఆ هड़౷ܭͱਪଌ౷ܭ ֬ɾ֬ม
֬ ظ ଟ࣍ݩͷ֬ม ඪຊநग़ αϯϓϦϯά -
ຊൃදͷ༰ େͷ๏ଇ ਖ਼ن த৺ۃݶఆཧ ύϥϝʔλʔͷਪఆͱ ࠷๏ ۠ؒਪఆ ౷ܭతݕఆ هड़౷ܭͱਪଌ౷ܭ ֬ɾ֬ม
֬ ظ ଟ࣍ݩͷ֬ม ඪຊநग़ αϯϓϦϯά -
هड़౷ܭ τʔλϧ ࠷ߴείΞ 990 ࠷είΞ 10 ฏۉείΞ 583.7 ඪ४ภࠩ 170.1
http://www.toeic.or.jp/toeic/about/data/data_avelist/data_ave01_04.html http://www.toeic.or.jp/toeic/about/data/data_avelist/data_dist01_04.html ྫɿ50&*$ެ։ςετୈճͷࢼݧ ˠୈճͷडݧऀશһΛରʹσʔλΛཁ͍ͨ͠ -
هड़౷ܭ τʔλϧ ࠷ߴείΞ 990 ࠷είΞ 10 ฏۉείΞ 583.7 ඪ४ภࠩ 170.1
ྫɿ50&*$ެ։ςετୈճͷࢼݧ ˠୈճͷडݧऀશһΛରʹσʔλΛཁ͍ͨ͠ http://www.toeic.or.jp/toeic/about/data/data_avelist/data_ave01_04.html http://www.toeic.or.jp/toeic/about/data/data_avelist/data_dist01_04.html खݩͷσʔλΛ ཁ͍ͯ͠Δɻ ੳͷରͱ͍ͨ͠ਓ શһͷσʔλ͕ಘΒΕ͍ͯΔ ͷͰɺͦΕΛ·ͱΊΔ -
ख࣋ͪͷσʔλΛαϚϦͯ͠Ѳ͢ΔͨΊͷख๏ هड़౷ܭ D = { x1, x2, · · ·
, xn } Oݸͷσʔλ FHOਓ͍Δͱ͋ΔΫϥεͷςετͷͳͲ جຊ౷ܭྔ ¯ x = 1 n n X i=1 xi ฏۉ 2 = 1 n n X i=1 ( xi ¯ x )2 ࢄ = v u u t 1 n n X i=1 ( xi ¯ x )2 ඪ४ภࠩ ˞ ਪଌΛߦ͏࣌ͷࢄ O ͰׂΔͱ͍͏ ͕͋Γ·͕ͦ͢Εޙ΄ͲͷϖʔδͰɻ -
ࢄɾඪ४ภࠩ ʮภࠩʯͱʁ *% ɹ ɹ ภࠩɹɹ
֤σʔλͷฏۉ͔Βͷࠩͷ͜ͱ -
ࢄɾඪ४ภࠩɿͷલʹฏۉภࠩ ͜ͷɺภࠩͷฏۉΛऔΓ͍͕ͨɾɾɾ *%ɹ ภࠩɹ ɹ
શ෦͢ͱ̌ʹͳͬͯ͠·͏ ࠨӈΓ߹͍͕औΕΔͱ͜Ζ͕ ฏۉͳͷͰ -
ࢄɾඪ४ภࠩɿͷલʹฏۉภࠩ *%ɹ ภࠩɹ ɹ ϓϥεԽ
స ͳͷͰɺϚΠφεΛऔͬͯΈΔ -
ࢄɾඪ४ภࠩɿͷલʹฏۉภࠩ *%ɹ ภࠩɹ ɹ ϓϥεԽ
ฏۉ ภࠩͷฏۉ ฏۉ͔ΒͷڑΛ ฏۉͨ͠ͷ ͳͷͰɺϚΠφεΛऔͬͯΈΔ -
ࢄɾඪ४ภࠩɿͷલʹฏۉภࠩ = 1 n n X i=1 | xi ¯
x | ฏۉภࠩ ภࠩΛ͠߹ΘͤͯσʔλOͰׂΔɻ ͭ·ΓฏۉΛͱ͍ͬͯΔɻ ภࠩΛશͯϓϥεʹม͑Δ ͨΊʹઈରΛऔΔɻ *%ɹ ภࠩɹ ɹ ϓϥεԽ ฏۉ -
ࢄɾඪ४ภࠩ ʴ ʴ ʴ ͍Ζ͍ΖͳαΠζͷਖ਼ํܗΛ ͯ͠ɺσʔλͷݸͰׂΔͱɺ ໘ੵͷฏۉ͕ٻ·Δɻ ͜Ε͕ࢄͷײతΠϝʔδɻ ʴʜʴ
ʴ ×ݸ ໘ੵͷฏۉ ᶃ ᶄ ᶅ ᶆ O -
ࢄɾඪ४ภࠩ ภࠩ = 1 n n X i=1 ( xi
¯ x )2 ࢄ ೋͨ͠ภࠩΛ͠߹ΘͤͯσʔλOͰׂΔɻ ͜ͷ߹ೋภࠩͷฏۉΛͱ͍ͬͯΔɻ ͭ·ΓɺΦϨϯδͷ໘ੵͷฏۉʹͳΔɻ ภࠩΛશͯϓϥεʹม͑ΔͨΊʹೋ͢Δɻ ʮೋ͢Δʯͱ͍͏͜ͱ໘ੵΛΠϝʔδͯ͠ྑ͍ɻ -
ࢄɾඪ४ภࠩ ඪ४ภࠩ = v u u t 1 N N
X i=1 ( xi ¯ x )2 ୯ҐΛ͞ʹͨ͠ ͞ ͞ ໘ੵ Y p ͞ ϧʔτ ໘ੵ͕͞ʹͲΔʂ ͷܭࢉΛ͢Δͱ -
ภࠩ ภࠩ ฏۉ ඪ४ภࠩ º ฏۉ ඪ४ภࠩ
º ฏۉͱͷภࠩΛඪ४ภࠩͰଌΓɺ̍ͭΛͱͯ͠ ฏۉΛʹͳΔΑ͏ௐͨ͠ͷɻ -
ਪଌ౷ܭ ྫɿ ڵຯͷ͋Δର ूஂ ͷ͕ͱͯଟ͍ɺ͘͠ ແݶʹଘࡏ͢Δ߹ɺඪຊநग़Λߦͬͯɺͦͷඪຊ ͔ΒूஂΛਪଌ͢Δɻ ΫϥεਓͳΒશһͷσʔλ͕ ಘΒΕΔͷͰهड़౷ܭͰे ຊਓਓஉੑສਓશһͷσʔλΛऔΔͷ
ඇৗʹίετ͕ߴ͍ͷͰɺҰ෦ͷσʔλ͔Β શମΛਪଌ͍ͨ͠ɻ -
ਪଌ౷ܭ ྫɿͰͷੜ࢈ɺҰ෦ͷܦࡁ౷ܭ ˠαϯϓϧ ඪຊ ͔Βશମ ूஂ Λਪଌ ਪଌ͢Δ ਪଌ͢Δ FYͷॏ͞
FYՈܭௐࠪ ਪଌʹ͏ෆ࣮֬ੑ Λ͏·͘ѻ͏ͨΊʹ ֬ͷ֓೦Λར༻͢Δ -
ຊൃදͷ༰ େͷ๏ଇ ਖ਼ن த৺ۃݶఆཧ ύϥϝʔλʔͷਪఆͱ ࠷๏ ۠ؒਪఆ ౷ܭతݕఆ هड़౷ܭͱਪଌ౷ܭ ֬ɾ֬ม
֬ ظ ଟ࣍ݩͷ֬ม ඪຊநग़ αϯϓϦϯά -
ࣄʹ͍ͭͯ ! 2 ⌦ = {!1, !2, · · ·
, !m } શࣄ ඪຊۭؒ ࣄ N௨Γͷ݁Ռ͕͋Δ ⌦ = { , } ཪ ද ྫɿίΠϯ͛ ද ཪ ࢼߦ͢Δͱ ! 2 { , } ཪ ද ͕ಘΒΕΔɻ ྫ͑ɺճ !(1) = ཪ !(2) = !(n) = ද ཪ ͳͲ ճ Oճ -
ྫɿαΠίϩ͛ ⌦ = {1, 2, 3, 4, 5, 6} ྫ͑ɺճ
!(1) = !(2) = !(n) = ճ Oճ ྫɿʢຊਓਓஉੑສਓʣ͔Β ɹɹϥϯμϜʹநग़ ճ ճ Oճ ⌦ = {!1, !2, · · · , !49870000 } ాத͞Μ !(1) = !43890298 = 171cm !(2) = !29184638 = 168cm ߴڮ͞Μ !(n) = !51398579 = 174cm ླ͞Μ 㱺 㱺 㱺 -
ྫɿๅ͘͡ ୈճֶΧϑΣๅ͘͡ *%
͋ͨΓ ୈճ !(1) = *% !(2) = *% *% ୈճ ୈճ !(n) = *% ୈճ ͋ͨΓɿສԁ ͣΕɿ̌ԁ !(3) = ͋ͨΓ 㱺ԁ 㱺ສԁ 㱺ԁ 㱺ԁ !1 !2 !3 !4 !5 !6 !7 !8 !9 !10 = {!1, !2, !3, · · · , !10 } ! 2 ⌦ = {ID1, ID2, ID3, · · · , ID10} -
֬มͱ ๅ͘͡ͷྫ X = X(!) ࣄΛ͋Δ࣮ʹରԠ͚ͨͷɻ ๅ͘͡ͷ߹બ ֹۚ ඪຊۭؒͷݩɹΛҾʹͱΓ࣮Λฦؔ͢ɻ ֬มࣗମʹෆنଇมಈཁૉͳ͘ɺͷબΕํ
ʹมಈཁૉ͕ଘࡏ͢Δɻ ⌦ ! ! X(!1) = 0 X(!2) = 0 X(!3) = 0 X(!4) = 0 X(!5) = 0 X(!6) = 0 X(!7) = 0 X(!8) = 0 X(!9) = 0 X(!10) = 100ສԁ ࣄʹֹۚΛ ׂΓ͍ͯͯΔ͚ͩ -
ๅ͘͡ͷྫ {! 2 ⌦ : X(!) 2 A} ୈճֶΧϑΣๅ͘͡ *%
͋ͨΓɿສԁ ͣΕɿ̌ԁ !1 !2 !3 !4 !5 !6 !7 !8 !9 !10 ͋ͨΓ ͋ͨΓ A X(!) = 100 ສԁͷू߹ Ac X(!) = 0 ԁͷू߹ ! 5 or ! 9 બֹ͕ۚສԁ Ͱ͋Δू߹ બֹۚ ݅ -
֬ PX (A) = P(X 2 A) = P({! 2
⌦ : X(!) 2 A}) ݅ʹͯ·ΔΑ͏ͳ શࣄͷݩશͯ ⌦ ݅ બֹۚສԁͱͳΔΑ͏ͳๅ͘͡ͰɺͦΕࠓճͷରͷશࣄ ͷதͰͳͷͰɺͦͷɺɹɹɹɹ͕ग़ݱ͢Δ֬ɻ !5, !9 !5, !9 શࣄ PX (A) = #({! 2 ⌦ : X(!) 2 A}) #( ) = #(!5, !9) #( ) = 2 10 = 0.2 શࣄ -
֬ͷఆٛ ίϧϞΰϩϑͷެཧ ඇෛੑɿɹͲΜͳࣄͷू߹"ʹରͯ͠ ֬ͷʹϚΠφεͳ͍ͷͰ શࣄͷ֬ɿ ɹɹɹɹɹ PX(⌦) =
1 શͯͷఆ͍ͯ͠ΔࣄͷͲΕ͔ ඞͣൃੜ͢Δ ՃࢉՃ๏ੑɿߴʑՃࢉݸͷू߹ྻ͕ ɹɹɹɹɹɹޓ͍ʹഉͳΒ A1, A2, · · · PX ([iAi) = X i PX (Ai) A1 A2 A3 A4 A5 A6 A7 A8 A9 A10 A11 A12 ʜ 0 PX(A) 1 -
ຊൃදͷ༰ େͷ๏ଇ ਖ਼ن த৺ۃݶఆཧ ύϥϝʔλʔͷਪఆͱ ࠷๏ ۠ؒਪఆ ౷ܭతݕఆ هड़౷ܭͱਪଌ౷ܭ ֬ɾ֬ม
֬ ظ ଟ࣍ݩͷ֬ม ඪຊநग़ αϯϓϦϯά -
֬ ֬ى͜Γ͢͞ͷࢦඪͰ͋ͬͨɻ·ͨɺ֬ม ɺඪຊ্ۭؒͷݩΛҾʹऔΓ࣮Λฦؔ͢ Ͱ͋ͬͨɻ ྫ͑ɺͱͳΔू߹ͱ͍ͬͨΑ͏ʹɺ֬ มͷʹΑΓάϧʔϐϯάΛ ͨ͠ͱ͖ʹɺͦͷΛ͑Δͱ X = X(!)
ඪຊۭؒ ⌦ ͱͳΔΑ͏ͳू߹ A A !1 !2 !3 !4 !5 !6 !7 !8 !11 !10 !9 !12 !13 !14 !15 !16 B C D X(!) = 0 X(!) = 0 #A = #{! 2 ⌦ : X(!) = 0} = 7 #B = #{! 2 ⌦ : X(!) = 1} = 2 #C = #{! 2 ⌦ : X(!) = 2} = 4 #D = #{! 2 ⌦ : X(!) = 3} = 3 -
֬ ඪຊۭؒ ⌦ ͱͳΔΑ͏ͳू߹ A A !1 !2 !3 !4
!5 !6 !7 !8 !11 !10 !9 !12 !13 !14 !15 !16 B C D X(!) = 0 ֬ 9 P(X = 0) = PX(A) = #{! 2 ⌦ : X(!) = 0} #⌦ = 7 16 P(X = 1) = PX (B) = #{! 2 ⌦ : X(!) = 1} #⌦ = 2 16 P(X = 2) = PX(C) = #{! 2 ⌦ : X(!) = 2} #⌦ = 4 16 P(X = 3) = PX(D) = #{! 2 ⌦ : X(!) = 3} #⌦ = 3 16 ֬มͷ͝ͱʹάϧʔϐϯάͯ֬͠ΛٻΊΔɻ -
֬ม9͕ͷΑ͏ʹLݸͷҟͳΔΛ ͱΔ߹ Λ9ͷ֬ͱݺͿɻ දهΛ؆୯ʹ͢ΔͨΊҎޙɺJΛলུ ֬ ֬ม͕ࢄͷ߹ { x1, x2, ·
· · , xk } P ( X = xi) = f ( xi) ֬ 9 ֬ 9 ֬ؔ ྦྷੵؔ F ( x ) = P ( X x ) ֬ม9͕͋ΔYҎԼͰ͋Δ֬ Λྦྷੵؔͱ͍͏ɻ -
֬ ֬ม͕࿈ଓͷ߹ ֬ม9͕͞ɺॏ͞ͳͲ࿈ଓ͢Δมͷ߹ɻ࿈ଓͷ ߹ P ( x < X
x + x ) x + x x ͷΑ͏ʹখ͍۠ؒ͞Λߟ͑ɺͦͷؒʹೖΔ֬Λߟ͑Δɻ ۠ؒͷ෯Ͱ͜ͷ֬Λׂͬͯɺͱͨ͠ۃݶΛ x x ! 0 f ( x ) = lim x !0 P ( x < X x + x ) x Λ֬ີؔͱݺͿɻ -
֬ ֬ม͕࿈ଓͷ߹ 9͕BͱCͷؒʹೖΔ֬Λੵͯ͠ x + x x f ( a
< x b ) = Z a b f ( x ) dx Ͱ༩͑ΒΕΔɻྦྷੵؔɺ f ( x ) F ( x ) = P ( X x ) = Z x 1 f ( u ) du ྦྷੵؔ ֬ີؔ -
ϕϧψʔΠ ද ཪ ࢼߦΛ̍ߦ͍ɺޭ͔ࣦഊͳͲ̎ ྫɿίΠϯ͛ ֬ؔ P ( X =
x ) = px(1 p )1 x ( x = 0 , 1) ύϥϝʔλʔ Qɿ ද ͕ग़Δ֬ ࢄ -
ϕϧψʔΠ # ϕϧψʔΠ͔ΒͷαϯϓϦϯάΛ࣮ߦ # ύϥϝʔλʔ p = 0.7 trial_size =
10000 set.seed(71) # ϕϧψʔΠʹै͏ཚੜ data <- rbern(trial_size, p) # ϕϧψʔΠͷ֬ dens <- data.frame(y=c((1-p),p)*trial_size, x=c(0, 1)) # άϥϑඳը ggplot() + layer(data=data.frame(x=data), mapping=aes(x=x), geom="bar", stat="bin", bandwidth=0.1 ) + layer(data=dens, mapping=aes(x=x, y=y), geom="bar", stat="identity", width=0.05, fill="#777799", alpha=0.7) 3ίʔυ ࢵɿཧతͳ֬ ࠇɿཚ͔Βੜͨ͠ώετάϥϜ -
ೋ߲ ϕϧψʔΠࢼߦΛOճ࣮ࢪ͠ɺԿճޭ͔ͨ͠ͷճ Λ֬มͱͨ͠ɻ ྫɿϑϦʔεϩʔ̍ηοτճ࣮ࢪɻ ɹɹ̍ճ͋ͨΓͷޭͰܭԿճೖ͔ͬͨ ֬ؔ ύϥϝʔλʔ Qɿ ද ͕ग़Δ֬<ϑϦʔεϩʔͷޭ֬>
P ( X = x ) = nCrpx(1 p )n x Oɿηοτ͋ͨΓͷճ ( x = 1 , 2 , · · · , n ) ࢄ -
ೋ߲ # ύϥϝʔλʔ p = 0.7 trial_size = 10000 sample_size
= 30 set.seed(71) # ϕϧψʔΠʹै͏ཚੜ gen_binom_var <- function() { return(sum(rbern(sample_size, p))) } result <- rdply(trial_size, gen_binom_var()) # ೋ߲ͷີؔ dens <- data.frame(y=dbinom(seq(sample_size), sample_size, 0.7))*trial_size # άϥϑඳը ggplot() + layer(data=resuylt, mapping=aes(x=V1), geom="bar", stat = "bin", binwidth=1, fill="#6666ee", color="gray" ) + layer(data=dens, mapping=aes(x=seq(sample_size)+.5, y=y), geom="line", stat="identity", position="identity",colour="red" ) + ggtitle("Bernoulli to Binomial.") ͷۂઢɿཧతͳ֬ ώετάϥϜɿཚ͔Βੜ 3ίʔυ -
ϙϫιϯ ηοτճ O ˠ 㱣 ɺ̍ճ͋ͨΓޭQ ˠ
OQΛҰఆͱͨ࣌͠ͷ ྫɿͷইͷνΣοΫΛߟ͑ΔɻDN෯Ͱইͷ༗ແΛ ΛௐΔɻϕϧψʔΠࢼߦͷ܁Γฦ͠ ˠ ೋ߲ ˠແݶʹׂͯ͠ϙΞιϯɻ DN DN DN DN DN DN DN DN DN DN DN શ෦Ͱই͕̓ͭ DN DN DN DN DN DN DN DN DN DN DN DN DN DN DN DN DN DN DN DN 㱣 ʹׂ ࢄ -
֬ؔ ύϥϝʔλʔ ͜ͷྫͷ߹ɺҰຊ͋ͨΓͷইͷฏۉతͳ Еɿฏۉੜىճ P ( X = x )
= e x x ! ( x = 1 , 2 , · · · , 1) ЕOQͷ͕ؔอͨΕ͍ͯΔ ̍ຊ͋ͨΓͷইͷ ͷˠϙΞιϯ ্ݶ͕ͳ͘ͳͬͨʂ ϙΞιϯ ࢄ -
ϙΞιϯ trial_size = 5000; width <- 1; # ͱͱͷઃఆ p
= 0.7; n = 10; np <- p*n # n!∞ɺp!0ɺnp=Ұఆ n = 100000; p <- np/n # ϕϧψʔΠʹै͏ཚੜ gen_binom_var <- function() { return(sum(rbern(n, p))) } result <- rdply(trial_size, gen_binom_var()) # ϙΞιϯͷີؔ dens <- data.frame(y=dpois(seq(20), np))*trial_size # άϥϑඳը ggplot() + layer(data=result, mapping=aes(x=V1), geom="bar", stat = "bin", binwidth=width, fill="#6666ee", color="gray" ) + layer(data=dens, mapping=aes(x=seq(20)+.5, y=y), geom="line", stat="identity", position="identity", colour="red" ) + ggtitle("Bernoulli to Poisson.") ͷۂઢɿཧతͳ֬ ώετάϥϜɿཚ͔Βੜ 3ίʔυ -
ਖ਼ن ೋ߲ͷOΛେ͖͘͢Δͱਖ਼نͰۙࣅͰ͖Δɻ ֬ີؔ ύϥϝʔλʔ Мɿඪ४ภࠩ Жɿฏۉ f(x) = 1 p
2⇡ 2 exp ⇢ 1 2 (x µ) 2 2 ( 1 < x < 1) ྫɿࢮ͵·ͰϑϦʔεϩʔ͛ͯܭԿճೖ͔ͬͨ ࿈ଓ -
ਖ਼ن ϙΞιϯͱͷҧ͍ DN DN DN DN DN DN DN DN
DN DN DN DN DN DN DN DN DN DN DN DN DN DN DN DN DN DN DN DN DN DN DN DN ʜ DN DN DN DN DN DN DN DN DN DN ˠ㱣DN DN DN DN DN DN DN DN DN DN Ԇʑͱ৳͢ ࿈ଓ -
ਖ਼ن # ύϥϝʔλʔ n <- 10000; p <- 0.7; trial_size
= 10000 width=10 # ϕϧψʔΠʹै͏ཚੜ gen_binom_var <- function() { return(sum(rbern(n, p))) } result <- rdply(trial_size, gen_binom_var()) # ਖ਼نͷີؔ dens <- data.frame(y=dnorm(seq(6800,7200), mean=n*p, sd=sqrt(n*p*(1-p)))*trial_size*width) # άϥϑඳը ggplot() + layer(data=result, mapping=aes(x=V1), geom="bar", stat = "bin", binwidth=width, fill="#6666ee", color="gray" ) + layer(data=dens, mapping=aes(x=seq(6800,7200), y=y), geom="line", stat="identity", position="identity", colour="red") + ggtitle("Bernoulli to Normal.") 3ίʔυ ͷۂઢɿཧతͳ֬ ώετάϥϜɿཚ͔Βੜ -
ඪ४ਖ਼ن ਖ਼نΛฏۉ̌ɺඪ४ภࠩ̍ʹͨ͠ͷ ֬ີؔ ύϥϝʔλʔ ͳ͠ ( 1 < x <
1) f(x) = 1 p 2⇡ exp ⇢ 1 2 x 2 ࿈ଓ -
ඪ४ਖ਼ن # ύϥϝʔλʔ n <- 10000; p <- 0.7 trial_size
= 30000 width=0.18 # ϕϧψʔΠʹै͏ཚੜ gen_binom_var <- function() { return(sum(rbern(n, p))) } result <- rdply(trial_size, gen_binom_var()) m <- mean(result$V1); sd <- sd(result$V1); result <- (result - m)/sd # ඪ४ਖ਼نͷີؔ dens <- data.frame(y=dnorm(seq(-4,4,0.05), mean=0, sd=1)*trial_size*width) # άϥϑඳը ggplot() + layer(data=result, mapping=aes(x=V1), geom="bar", stat = "bin", binwidth=width, fill="#6666ee", color="gray" ) + layer(data=dens, mapping=aes(x=seq(-4,4,0.05), y=y), geom="line", stat="identity", position=“identity", colour="red" ) + ggtitle("Bernoulli to Standard Normal.") 3ίʔυ ͷۂઢɿཧతͳ֬ ώετάϥϜɿཚ͔Βੜ -
ΧΠೋ ඪ४ਖ਼نʹै͏֬มΛ̎ͯ͠Lݸ ͠߹Θͤͨ ɹɹɹɹɹɹɹɹɹɹɹɹɹɹɹɹɹ͕ै͏ ֬ີؔ ύϥϝʔλʔ Lɿࣗ༝ ̎ͨ͠ඪ४ਖ਼نͷ֬มͷ ɹɹɹɹɹ͠߹Θͤͨ f
( x, k ) = (1 / 2)k/ 2 ( k/ 2) xk/ 2 1 e x/ 2 (0 x ) Xi Z = X2 1 + · · · + X2 k ࿈ଓ -
ฏۉΛʹͣΒ͢ Oճ࣮ࢪ ίΠϯ͛ දͷ֬Q Oˠ㱣 ΧΠೋ ฏۉΛʹͣΒ͢ ඪ४ภࠩΛʹ ॖΊΔ ฏۉΛʹͣΒ͢
Lͷͱ͖ ೋͨ͠ͱ͜Ζ ʹੵΜͰ͍͘ Lͷͱ͖ Lͭ αϯϓϦϯά͢Δ Lͭͷ͞Λ ͨ͠ͷΛ ϓϩοτ͢Δ ࿈ଓ -
ΧΠೋ # ύϥϝʔλʔ p <- 0.7; n <- 1000; trial_size
<- 100000; width <- 0.3; df <- 3 # ϕϧψʔΠʹै͏ཚੜ(3·Θ͠) gen_binom_var <- function() { return(sum(rbern(n, p))) } gen_chisq_var <- function() { result <- rdply(trial_size, gen_binom_var()) return(((result$V1 - mean(result$V1))/sd(result$V1))**2) } # ࣗ༝dfͷ͚ͩੜ͢Δ result <- rlply(df, gen_chisq_var(),.progress = "text") res <- data.frame(x=result[[1]] + result[[2]] + result[[3]]) # ΧΠೋͷີؔ(ࣗ༝=3) xx <- seq(0,20,0.1) dens <- data.frame(y=dchisq(x=xx, df=df)*trial_size*width) # άϥϑඳը ggplot() + layer(data=data, mapping=aes(x=x), geom="bar", stat = "bin", binwidth=width, fill="#6666ee", color="gray" ) + layer(data=dens, mapping=aes(x=xx, y=y), geom="line", stat="identity", position="identity", colour="blue" ) + ggtitle("Bernoulli to Chisquare") 3ίʔυ ͷۂઢɿཧతͳ֬ ώετάϥϜɿཚ͔Βੜ -
ࢦ ϙΞιϯ͕୯Ґʢ͘͠୯Ґ࣌ؒʣʹฏۉ Կճੜى͢Δ͔ɺΛද͢Ͱ͕͋ͬͨɺಉ͡ࣄ Λҧ͏֯Ͱଊ͑ͨ͠ͷ͕ࢦɻ DN DN DN DN DN DN
DN DN DN DN DN શ෦Ͱই͕̓ͭ ϙΞιϯతߟ͑ํ ࢦతߟ͑ํ ࣍ൃੜ͢Δ·ͰʹɺͲͷ͘Β͍ͷڑ ࣌ؒ ͔ ࿈ଓ ࠓεΩοϓ -
ࢦ ֬ີؔ ύϥϝʔλʔ Еɿ୯Ґʢ࣌ؒʣʹฏۉతʹੜى͢Δճ f ( x, ) = ⇢
e x ( x 0) 0 ( x < 0) ࿈ଓ -
trial_size = 7000; width <- .01; # ͱͱͷઃఆ p =
0.7; n = 10; np <- p*n; # n!∞ɺp!0ɺnp=Ұఆ n = 10000; p <- np/n # ϕϧψʔΠʹै͏ཚੜ gen_exp_var <- function() { cnt <- 0 while (TRUE) { cnt <- cnt + 1 if (rbern(1, p)==1){ return(cnt) # 1͕ग़ͨΒͦΕ͕Կճ͔Λฦ͢ } } } data <- data.frame(x=rdply(trial_size, gen_exp_var())/n) names(data) <- c("n", "x") # ࢦͷີؔ dens <- data.frame(y=dexp(seq(0, 1.5, 0.1), np)*trial_size*width) ggplot() + layer(data=data, mapping=aes(x=x), geom="bar", stat = "bin", binwidth=width, fill="#6666ee", color="gray" ) + layer(data=dens, mapping=aes(x=seq(0, 1.5, 0.1), y=y), geom="line", stat="identity", position="identity", colour="red" ) + ggtitle("Bernoulli to Exponential.") 3ίʔυ ͷۂઢɿཧతͳ֬ ώετάϥϜɿཚ͔Βੜ ࢦ -
ΨϯϚ ֬ີؔ ύϥϝʔλʔ ୯Ґʢ࣌ؒʣ͋ͨΓЌճൃੜ͢Δࢦͷ Ћݸͷͷ֬Λߟ͑ͨ࣌ɺΨϯϚͱͳΔ f(x, ↵, ) = ↵
(↵) x ↵ 1 exp( x) (0 x < 1) ↵ X i=1 Xi ⇠ (↵, ) Xi ⇠ Exp( ) ࿈ଓ ࠓεΩοϓ -
trial_size = 7000; width <- .035; # ͱͱͷઃఆ p =
0.7; n = 10; np <- p*n; # n!∞ɺp!0ɺnp=Ұఆ n = 10000; p <- np/n; alpha <- 5 # ϕϧψʔΠʹै͏ཚੜ get_interval <- function(){ cnt <- 0 while (TRUE) { cnt <- cnt + 1 if (rbern(1, p)==1){ return(cnt) } } } gen_exp_var <- function() { data <- data.frame(x=rdply(trial_size, get_interval())/n) names(data) <- c("n", "x") return(data) } result <- rlply(alpha, gen_exp_var()) data <- data.frame(x=result[[1]]$x + result[[2]]$x + result[[3]]$x + result[[4]]$x + result[[5]]$x) # ΨϯϚͷີؔ dens <- data.frame(y=dgamma(seq(0, 3,.01), shape=alpha, rate=np)*trial_size*width) ggplot() + layer(data=data, mapping=aes(x=x), geom="bar", stat = "bin", binwidth=width, fill="#6666ee", color="gray" ) + layer(data=dens, mapping=aes(x=seq(0,3,.01), y=y), geom="line", stat="identity", position="identity", colour="red" ) + ggtitle("Bernoulli to Gamma") 3ίʔυ ͷۂઢɿཧతͳ֬ ώετάϥϜɿཚ͔Βੜ ΨϯϚ -
ٯΨϯϚ ֬ີؔ ύϥϝʔλʔ f(x, ↵, ) = ↵ (↵) x
(↵+1) exp ✓ x ◆ (0 x < 1) ୯Ґʢ࣌ؒʣ͋ͨΓЌճൃੜ͢ΔࢦͷЋݸͷͷ ֬Λߟ͑ͨ࣌ɺΨϯϚͱͳΔ Xi ⇠ Exp( ) Z = ↵ X i=1 Xi ⇠ (↵, ) 1/Z ⇠ IG(↵, ) ͜ͷ࣌ɺ;ͷٯ͕ै͏ΛٯΨϯϚͱݴ͏ɻ ࿈ଓ ࠓεΩοϓ -
ฏۉΛʹͣΒ͢ Oճ࣮ࢪ ίΠϯ͛ දͷ֬Q OQҰఆͰOˠ㱣 ฏۉΛʹͣΒ͢ ִؒͷ͞Λϓϩοτ ٯΨϯϚ Lͭͷ͞Λ ͨ͠ͷΛ
ϓϩοτ͢Δ Lͭ αϯϓϦϯά શͯͷσʔλΛ ZYͰม [ [ X X ࿈ଓ -
trial_size = 7000; width <- .; # ͱͱͷઃఆ p =
0.7; n = 10; np <- p*n; # n!∞ɺp!0ɺnp=Ұఆ n = 10000; p <- np/n; alpha <- 5 # ϕϧψʔΠʹै͏ཚੜ get_interval <- function(){ cnt <- 0 while (TRUE) { cnt <- cnt + 1 if (rbern(1, p)==1){ return(cnt) } } } gen_exp_var <- function() { data <- data.frame(x=rdply(trial_size, get_interval())/n) names(data) <- c("n", "x") return(data) } result <- rlply(alpha, gen_exp_var()) data <- data.frame(x=1/(result[[1]]$x + result[[2]]$x + result[[3]]$x + result[[4]]$x + result[[5]]$x)) # ٯΨϯϚͷີؔ dens <- data.frame(y=dinvgamma(seq(0, 23,.01), shape=5, rate=1/np)*trial_size*width) ggplot() + layer(data=data, mapping=aes(x=x), geom="bar", stat = "bin", binwidth=width, fill="#6666ee", color="gray" ) + layer(data=dens, mapping=aes(x=seq(0,3,.01), y=y), geom="line", stat="identity", position="identity", colour="red" ) + ggtitle("Bernoulli to Inversegamma") 3ίʔυ ͷۂઢɿཧతͳ֬ ώετάϥϜɿཚ͔Βੜ ٯΨϯϚ -
ඪ४Ұ༷ ֬ີؔ ύϥϝʔλʔ ͳ͠ ͔ΒͷؒͰ֬Ͱൃੜ͢ΔΑ͏ͳݱΛ ද͢ f(x) = ⇢ 1
(0 x 1) 0 (otherwise) ࿈ଓ ࠓεΩοϓ -
ඪ४Ұ༷ ίΠϯ͛ දͷ֬Q Rճ࣮ࢪ ࢼߦ݁ՌΛ֤ܻʹׂΓͯ ද ཪ ද ཪ ද
ཪ ճ ճ Rճ ˞ͬͱޮͷྑ͍Γํ͋Δͱࢥ͍·͕͔͢Γ͢͞ͷͨΊʜ Z = x1(1 / 2)1 + x2(1 / 2)2 + · · · + xq(1 / 2)q ࿈ଓ -
width <- 0.02 p <- 0.5; sample_size <- 1000 trial_size
<- 100000 gen_unif_rand <- function() { # sample_sizeܻͷ2ਐগΛϕϧψʔΠʹ # ै͏ཚ͔Βੜ return (sum(rbern(sample_size, p) * (rep(1/2, sample_size) ** seq(sample_size)))) } gen_rand <- function(){ return( rdply(trial_size, gen_unif_rand()) ) } system.time(res <- gen_rand()) ggplot() + layer(data=res, mapping=aes(x=V1), geom="bar", stat = "bin", binwidth=width, fill="#6666ee", color="gray" ) + ggtitle("Bernoulli to Standard Uniform") 3ίʔυ ͷۂઢɿཧతͳ֬ ώετάϥϜɿཚ͔Βੜ ඪ४Ұ༷ -
֬ີؔ ύϥϝʔλʔ BɿԼݶ Cɿ্ݶ f(x, a, b) = ⇢ (b
a) 1 (a x b) 0 (otherwise) B͔ΒCͷؒͰ֬Ͱൃੜ͢ΔΑ͏ͳݱΛ ද͢ Ұ༷ ࿈ଓ ࠓεΩοϓ -
a <- 5 b <- 8; width <- 0.05 p
<- 0.5 sample_size <- 1000 trial_size <- 500000 gen_unif_rand <- function() { # sample_sizeܻͷ2ਐগΛϕϧψʔΠʹ # ै͏ཚ͔Βੜ return (sum(rbern(sample_size, p) * (rep(1/2, sample_size) ** seq(sample_size)))) } gen_rand <- function(){ return( rdply(trial_size, gen_unif_rand()) ) } system.time(res <- gen_rand()) res$V1 <- res$V1 * (b-a) + a ggplot() + layer(data=res, mapping=aes(x=V1), geom="bar", stat = "bin", binwidth=width, fill="#6666ee", color="gray" ) + ggtitle("Bernoulli to Uniform") + xlim(4,9) 3ίʔυ ͷۂઢɿཧతͳ֬ ώετάϥϜɿཚ͔Βੜ Ұ༷ -
ϕʔλ ֬ີؔ ύϥϝʔλʔ Ћɿฒସ͑ͨ͋ͱͷαϯϓϦϯά͢Δϙδγϣϯ Ќɿಉ͘͡ϙδγϣϯ͕ͩɺޙΖ͔Β͑ͨͱ͖ͷͷ f ( x, ↵, )
= 1 B ( ↵, )x ↵ 1(1 x ) 1 (0 < x < 1) ಠཱʹඪ४Ұ༷ʹै͏ Ћ Ќ ݸͷ֬ม Λେ͖͞ͷॱʹฒସ͑ͨͱ͖ɺখ͍͞ํ͔Β Ћ൪Ίʢେ͖͍ํ͔ΒЌ൪ʣͷ֬ม9ͷ͕ ϕʔλ# Ћ Ќ ͱͳΔɻ ˠ# ɾ ϕʔλؔ Xi ⇠ U(0, 1) iid (i = 1, 2, · · · , ↵ + 1) ʜ ʜ Ћݸ Ќݸ ࿈ଓ ࠓεΩοϓ -
width <- 0.03; p <- 0.5 digits_length <- 30; set_size
<- 3 trial_size <- 30000 gen_unif_rand <- function() { # digits_lengthܻͷ2ਐগΛϕϧψʔΠ # ʹै͏ཚ͔Βੜ return (sum(rbern(digits_length, p) * (rep(1/2, digits_length) ** seq(digits_length)))) } gen_rand <- function(){ return( rdply(set_size, gen_unif_rand())$V1 ) } unif_dataset <- rlply(trial_size, gen_rand, .progress='text') p <- ceiling(set_size * 0.5); q <- set_size - p + 1 get_nth_data <- function(a){ return(a[order(a)][p]) } disp_data <- data.frame(lapply(unif_dataset, get_nth_data)) names(disp_data) <- seq(length(disp_data)); disp_data <- data.frame(t(disp_data)) names(disp_data) <- "V1" x_range <- seq(0, 1, 0.001) dens <- data.frame(y=dbeta(x_range, p, q)*trial_size*width) ggplot() + layer(data=disp_data, mapping=aes(x=V1), geom="bar", stat = "bin", binwidth=width, fill="#6666ee", color="gray" ) + layer(data=dens, mapping=aes(x=x_range, y=y), geom="line", stat="identity", position="identity", colour="red" ) + ggtitle("Bernoulli to Beta") 3ίʔυ ͷۂઢɿཧతͳ֬ ώετάϥϜɿཚ͔Βੜ ϕʔλ -
ຊൃදͷ༰ େͷ๏ଇ ਖ਼ن த৺ۃݶఆཧ ύϥϝʔλʔͷਪఆͱ ࠷๏ ۠ؒਪఆ ౷ܭతݕఆ هड़౷ܭͱਪଌ౷ܭ ֬ɾ֬ม
֬ ظ ଟ࣍ݩͷ֬ม ඪຊநग़ αϯϓϦϯά -
E[X] = X( )P( ) + X( )P( ) =
0 ⇥ 0.8 + 1, 000, 000 ⇥ 0.2 = 200, 000 ظ ͣΕԁ Q ͋ͨΓສԁ ֬มͷͱͦͷ֬Λֻ͚͋Θ͑ͤͯͨͨ͠ͷɻ ฏۉతʹಘΒΕΔͱղऍͰ͖Δɻ ͣΕ ͋ͨΓ ͋ͨΓ E[ X ] = X x xp ( x ) ͭ·Γ ͷΑ͏ʹఆٛ͞ΕΔɻͰදݱ ͞ΕΔ͜ͱଟ͍ɻ ֬ม ͷ ֬ µ ͣΕ -
ظ ೋ߲ͷྫ O Yͷೋ߲ ೋ߲ ✓ n x ◆
= n ! ( n x )! x ! E[ X ] = n X x =0 xP ( x ) = n X x =0 x ✓ n x ◆ px(1 p )n x = n X x =0 x n ! ( n x )! x !px(1 p )n x = n X x =0 n ( n 1)! ( n x )!( x 1)!px(1 p )n x = np n X x =0 ✓ n 1 m 1 ◆ p ( x 1)(1 p )( n 1) ( x 1) = np -
ظ Var[ X ] = E[( X E [ X
])2] = X x ( x E[ x ])2 P ( x ) = 2 ɹ͔Βͷภࠩʹରͯ͠ɺ֬ͷΛॏΈͱͯ͠ Ճॏฏۉͨ͠ͷɻࢄΒΓͷईɻ µ -
Var[ x ] = E[( X E[ X ])2] =
Z 1 1 ( x E [ x ])2 f ( x ) dx = 2 ظ ࿈ଓͷ֬มͷ߹ ظ ࢄ E[ X ] = Z 1 1 xf ( x ) dx = µ -
ظ ֬มΛؔͰม͔ͯ͠ΒظΛͱΔ͜ͱྑ͘ ߦΘΕΔɻ E[ g ( X )] = Z
1 1 g ( x ) f ( x ) dx g(X) = (X E[X])2 ɹɹɹɹɹɹɹɹɹɹͱ͓͘ͱɺࢄ͕දݱͰ͖Δɻ ظΦϖϨʔλʔ E[ · ] = Z 1 1 · f ( x ) dx -
ɹɹɹɹɹͱͨ࣌͠ͷظɺ ظ ੵ g ( x ) = x k
E[ g ( X )] = E[ X k] = Z 1 1 x k f ( x ) dx ɺݪ·ΘΓͷL࣍ͷੵͱݺΕΔɻ ·ͨͱ͋ΒΘ͢ɻ ͳͷͰɺݪ·ΘΓͷ࣍ͷੵͱฏۉͷ͜ͱͱͳΔɻ µ0 k -
ظ ࣍ʹɺɹɹɹɹɹͱ͢Δͱ ੵ ɺฏۉ·ΘΓͷL࣍ͷੵͱݺΕɺͱ͋ΒΘ͢ɻ ͳͷͰɺฏۉ·ΘΓͷ࣍ͷੵͱࢄͷ͜ͱͱͳΔɻ g ( x ) =
( x E [ x ])k E[ g ( X )] = E[( X E[ X ]])k] = Z 1 1 ( x E[ x ])k f ( x ) dx µk -
ظ ޙͰ͏ܭࢉ E[ cX ] = c E[ X ]
* E[ cX ] = Z 1 1 cxf ( x ) dx = c Z 1 1 xf ( x ) dx = c E[ X ] -
ظ Var[ cX ] = c 2Var[ X ] *
Var[ cX ] = Z 1 1 ( cx E[ cx ])2 f ( x ) dx = Z 1 1 ( cx cµ )2 f ( x ) dx = Z 1 1 c 2( x µ )2 f ( x ) dx = c 2 Z 1 1 ( x µ )2 f ( x ) dx = c 2Var[ X ] ޙͰ͏ܭࢉ -
ຊൃදͷ༰ େͷ๏ଇ ਖ਼ن த৺ۃݶఆཧ ύϥϝʔλʔͷਪఆͱ ࠷๏ ۠ؒਪఆ ౷ܭతݕఆ هड़౷ܭͱਪଌ౷ܭ ֬ɾ֬ม
֬ ظ ଟ࣍ݩͷ֬ม ඪຊநग़ αϯϓϦϯά -
ଟ࣍ݩͷ֬ม ̍ͭͷ֬มɺ̍ͭͷσʔλʹ૬͢Δɻ ݱ࣮ͷσʔλੳͰଟͷσʔλ͔ΒਪଌΛߦ͏ͷͰ ଟ࣍ݩͷ֬มͷཧղ͕ॏཁͱͳΔɻ ·ͣ̎࣍ݩͷ߹͔Βɻ ͜͜ʹೖΔ֬ Λߟ͑ͱ͢Δͱ P ( x
< X 5 x + x, y < Y 5 y + y ) x, y ! 0 f ( x, y ) = lim x, y !0 P ( x < X 5 x + x, y < Y 5 y + y ) ͜ͷɹɹɹɹΛಉ࣌֬ີؔͱ ɹɹɹɹɹɹɹɹɹɹɹɹɹɹݺͿ f ( x, y ) -
ଟ࣍ݩͷ֬ม g ( x ) = Z 1 1 f
( x, y ) dy h ( y ) = Z 1 1 f ( x, y ) dx g ( x ) h(y) पลԽ ม ม ม ม YͱZͷݸʑͷ֬ΛٻΊΔ͜ͱΛपลԽͱ͍͏ ΛٻΊ͍ͨมҎ֎ͷ มʹ͍ͭͯੵͯ͠ٻΊΔɻ ͬͪ͜ʹͭͿ͢ ͬͪ͜ʹ ͭͿ͢ -
ଟ࣍ݩͷ֬ม EX,Y [ g ( X, Y )] = Z
1 1 Z 1 1 g ( x, y ) f ( x, y ) dxdy ̎ͭͷ֬มͷؔͷظ ྫ g ( x, y ) = x 0.8 y 0.8 ( x, y ) ⇠ N ((4 , 4) , S ) S = 1 0.5 0.4 1 EX,Y [ g(X, Y )] = 8.02 㱺 -
ଟ࣍ݩͷ֬ม g(X, Y ) = (X µX)(Y µY ) ڞࢄ
ͱͯ͠ظΛͱΔͱ ڞࢄ Cov[ X, Y ] = E [( X µX)( Y µY )] ͕ಘΒΕΔɻ -
g(X, Y ) = (X µX)(Y µY ) ϚΠφε ϚΠφε
ϓϥε ϓϥε µX ϚΠφε ϚΠφε ϓϥε ϓϥε µX ϚΠφε ϚΠφε ϓϥε ϓϥε ϚΠφε ϚΠφε ϓϥε ϓϥε µX µX µY µY µY µY ڞࢄͷײతཧղ S1 = S2 = S3 = S4 = 1 0.8 0.8 1 1 0.8 0.8 1 1 0 0 1 1 0.999 0.999 1 Cov[ X, Y ] = E [( X µX)( Y µY )] ( x, y ) ⇠ N ((4 , 4) , S ) -
ଟ࣍ݩͷ֬ม ಠཱͳ֬ม ಉ࣌֬ɹɹɹɹΛ f ( x, y ) f (
x, y ) = g ( x ) h ( y ) ͷΑ͏ʹ̎ͭͷपล֬ͷੵͰදͤΔ࣌ɺ ɹɹYͱZಠཱͳ֬มͰ͋Δ ͱ͍͏ɻ -
ଟ࣍ݩͷ֬ม ಠཱͳ֬มͷڞࢄ Cov[X, Y ] = E [(X µX)(Y µY
)] = Z 1 1 Z 1 1 (x µX )(y µy)f(x, y)dxdy = Z 1 1 (x µX)g(x)dx Z 1 1 (y µy)h(y)dy f ( x, y ) = g ( x ) h ( y ) ฏۉपΓͷ࣍ͷੵ ฏۉपΓͷ࣍ͷੵ = 0 -
ଟ࣍ݩͷ֬ม ʮͭʯ͔Βʮଟʯͷ֦ு ( x1, x2, · · · , xn)
ͷOݸͷ֬ม͕͋Δ࣌ɺ ɹͷपล֬ɺҎ֎ͷ֬มશͯͰੵͯ͠ x1 f ( x1) = Z · · · Z f ( x1, · · · , xn) dx2 · · · dxn x1 ಉ࣌֬ີ ·ͨɺ f ( x1, · · · , xn) = f ( x1) · · · f ( xn) ͕Γཱͭ࣌ɺɹɹɹɹɹޓ͍ʹಠཱͰ͋Δɺͱ͍͏ɻ x1 · · · xn -
ఆཧɿɹɹɹɹɹޓ͍ʹಠཱͳ֬มͱ͢Δɻɹɹɹɹ ɹɹɹɹɹɹɹɹ͕ͦΕͧΕɹɹɹɹɹͷΈͷؔͷ࣌ɺ ଟ࣍ݩͷ֬ม x1 · · · xn g1( x1)
, · · · , gn( xn) x1 · · · xn E[ n Y i=1 gi( xi)] = n Y i=1 E[ gi( xi)] ͕Γཱͭɻ ֻ͚ࢉͷظ ظͷֻ͚ࢉʹग़དྷΔ ূ໌ɿ E[ g1( x1)] E[ gn( xn)] E[ n Y i=1 gi( xi)] = Z 1 1 · · · Z 1 1 g1( x1) · · · gn( xn) f ( x1, · · · , xn) dx1 · · · dxn = Z 1 1 g1( x1) f ( x1) dx1 · · · Z 1 1 gn( xn) f ( xn) dxn = n Y i=1 E[ gi( xi)] f ( x1) · · · f ( xn) ಠཱͳͷͰ -
ଟ࣍ݩͷ֬ม ఆཧɿɹɹɹɹɹޓ͍ʹಠཱͳ֬มͱ͢Δɻ ɹฏۉɹɺࢄɹΛ࣋ͭͱ͢Δɻ ɹɹɹɹɹɹ ͋Δఆɹɹɹɹɹɹɹ͕͋Δͱ͖ɺɹɹɹɹɹɹɹ ฏۉ ࢄ Λ࣋ͭ֬มͱͳΔɻ x1
· · · xn xi µi 2 i i = 1, 2, · · · , n c = (c1, · · · , cn) c1x1 + · · · + cnxn c1µ1 + · · · + cnµn c2 1 2 1 + · · · + c2 n 2 n -
ଟ࣍ݩͷ֬ม ূ໌ɿ ฏۉ E[ c1x1 + · · · +
cnxn] = Z 1 1 · · · Z 1 1 ( c1x1 + · · · + cnxn) f ( x1 · · · , xn) dx1 · · · dxn = c1 Z 1 1 · · · Z 1 1 x1f ( x1 · · · , xn) dx1 · · · dxn · · · cn Z 1 1 · · · Z 1 1 xnf ( x1 · · · , xn) dx1 · · · dxn = c1 Z 1 1 x1dx1 · · · cn Z 1 1 xndxn = c1µ1 + · · · + cnµn f ( x1) · · · f ( xn) f ( x1) · · · f ( xn) µ1 µn = c1 Z 1 1 x1dx1 · · · cn Z 1 1 xndxn = c1µ1 + · · · + cnµn -
ଟ࣍ݩͷ֬ม ূ໌ɿ ࢄ Var[ c1x1 + · · · +
cnxn] = E[{( c1x1 + · · · + cnxn) E[ c1x1 + · · · + cnxn]}2] = E[{ c1( x1 µ1) + · · · + c1( x1 µ1)}2] = E[ n X i=1 c 2 i ( xi µi)2 + X i6=j cicj( xi µj)( xi µj)] = n X i=1 c 2 i E[( xi µi)2] + X i6=j cicjE[( xi µj)( xi µj)] = c 2 1 2 1 + · · · + c 2 n 2 n c1µ1 + · · · + cnµn = E[ xi µi]E[ xj µj] = 0 ಠཱͳͷͰ = 2 i -
ຊൃදͷ༰ େͷ๏ଇ ਖ਼ن த৺ۃݶఆཧ ύϥϝʔλʔͷਪఆͱ ࠷๏ ۠ؒਪఆ ౷ܭతݕఆ هड़౷ܭͱਪଌ౷ܭ ֬ɾ֬ม
֬ ظ ଟ࣍ݩͷ֬ม ඪຊநग़ αϯϓϦϯά -
ɹɹɹɹɺޓ͍ʹಠཱ͔ͭɺಉҰͷ͔Βͷ֬ม ͱ͢Δɻ ͜ͷޓ͍ʹಠཱ͔ͭಉҰͷͰ͋Δ͜ͱΛ JJE*OEFQFOEFOUBOE*EFOUJDBMMZ%JTUSJCVUFEͱ͍͏ɻ ඪຊநग़ αϯϓϦϯά αϯϓϦϯάͨ֬͠มͷ ඪຊ நग़ϓϩηε ूஂ
ਪଌϓϩηε x1 · · · xn ͜ͷͱ͖ɺOݸͷ֬มͷྻͷ͜ͱΛ ֬ඪຊͱݺͿɻ x1 · · · xn xi ूஂ͔Βແ࡞ҝநग़͞ΕͨͷͰ͋Γɺूஂͷฏ ۉΛฏۉɺूஂͷࢄɹΛࢄͱ͍͏ɻ µ 2 (µ, 2) -
ඪຊநग़ αϯϓϦϯά ඪຊͷฏۉɺࢄ ಉ༷ʹɹɹɹɹɹΛJJEͰ͋Δ֬มͱ͠ɺͦͷ x1 · · · xn T
= x1 + · · · + xn Λඪຊͱఆٛ͢Δɻ ฏۉ E[ T ] = E[ x1 + · · · + xn] = E[ x1] + · · · + E[ xn] = nµ ࢄ Var[ T ] = Var[ x1 + · · · + xn] = Var[ x1] + · · · + Var[ xn] = n 2 2 1 = · · · = 2 n c1 = · · · = cn = 1 Var[ c1x1 + · · · + cnxn] = c 2 1 2 1 + · · · + c 2 n 2 n ͔ͭ ͷͱ͖ͷ Λར༻ -
ࢄࢄɹΛOͰׂͬͨͷʹͳΔɻͭ·Γɺ ࢄαϯϓϧαΠζΛ૿ͤ૿͢΄Ͳখ͘͞ͳΔʂ ඪຊநग़ αϯϓϦϯά ͜͜Ͱඪຊฏۉ ¯ x = 1 n
n X i=1 xi = 1 n T ͷظʢฏۉʣ E[¯ x ] = 1 n E[ T ] = n · 1 n µ = µ ͷΑ͏ʹฏۉɹͱҰக͠ɺ Var[¯ x ] = Var[ 1 n T ] = 1 n 2 Var[ T ] = 2 n ॏཁʂ ॏཁʂ µ 2 -
ඪຊநग़ αϯϓϦϯά < >ͷҰ༷ཚΛݸੜͯ͠ฏۉΛࢉग़ɻ͜ΕΛ ճτϥΠΞϧ͠ɺ֬ೝͯ͠ΈΔɻ ࠷ॳͷճͷτϥΠΞϧ ͷτϥΠΞϧͰಘΒΕͨ ฏۉͷώετάϥϜ ฏۉ ࢄ
͕ಘΒΕ͓ͯΓɺظ Ͱ͋Δ Var[¯ x ] = 2 n = 0 . 0833 500 = 0 . 000166 E[¯ x ] = 0 . 5 ͱҰக͍ͯ͠Δɻ -
ຊൃදͷ༰ େͷ๏ଇ ਖ਼ن த৺ۃݶఆཧ ύϥϝʔλʔͷਪఆͱ ࠷๏ ۠ؒਪఆ ౷ܭతݕఆ هड़౷ܭͱਪଌ౷ܭ ֬ɾ֬ม
֬ ظ ଟ࣍ݩͷ֬ม ඪຊநग़ αϯϓϦϯά -
େͷ๏ଇ νΣϏγΣϑͷෆࣜ YΛ༗ݶͷฏۉͱࢄΛ࣋ͬͨ֬มͱ͢Δɻ ͜ͷͱ͖ ɹɹɹɹɹɹɹɹɹɹɹɹɹɹɹɹɹɹɹ͕Γཱͭɻ ʢYͷ͕ͲͷΑ͏ͳͷͰ͋ͬͯ༗ݶͷɹɺɹΛ࣋ͯΓཱͭʣ µ 2 P (|
x µ | > ) 5 1 2 µ 2 ЕМ͕YЖͷ͞ΑΓ ͘ͳΔ֬ҎԼ 1/ 2 = 1 ) P (| x µ | > ) 5 1 ΞλϦϚΤ = 2 ) P (| x µ | > ) 5 1 / 4 = 3 ) P (| x µ | > ) 5 1 / 9 ʜ -
େͷ๏ଇ νΣϏγΣϑͷෆࣜɿূ໌ 2 = Z 1 1 ( x µ
)2 f ( x ) dx = Z I1 ( x µ )2 f ( x ) dx + Z I2 ( x µ )2 f ( x ) dx + Z I3 ( x µ )2 f ( x ) dx 㱢ͳͷͰ͜ΕΛফͯ͠ෆࣜʹ͢Δɻ 2 = Z I1 ( x µ )2 f ( x ) dx + Z I3 ( x µ )2 f ( x ) dx = Z I1 2 2 f ( x ) dx + Z I3 2 2 f ( x ) dx = 2 2[ P ( x 2 I1) + P ( x 2 I3)] ͱ͢Δͱɺ I1 = ( 1, µ ), I2 = [µ , µ + ], I3 = (µ + , 1) = P (| x µ | > ) P (| x µ | > ) 5 1 2 ) -
ͱ͢Δɻҙͷʹରͯ͠ ɹɹɹɹɹΛJJEͰ͋Δ֬มͱ͠ɺ༗ݶͷฏۉɹɺ ࢄɹͷʹै͏ͱ͢ΔɻOݸͷσʔλͷඪຊฏۉΛ େͷ๏ଇ ඪຊฏۉʹର͢Δେͷ๏ଇʢऑ๏ଇʣ x1 · · · xn
µ 2 " > 0 ͕Γཱͭɻ͜ΕΛඪຊฏۉฏۉʹ֬ऩଋ͢Δɺ ͱ͍͍ lim n!1 P {|¯ xn µ | = " } = 0 ¯ xn = 1 n n X i=1 xi ¯ xn µ ¯ xn ! µ in P ͱ͔͘ɻ -
େͷ๏ଇ ඪຊฏۉʹର͢Δେͷ๏ଇʢऑ๏ଇʣɿূ໌ ҙͷʹରͯ͠ " > 0 Џ͕খ͘͞ͱOΛ૿ͤ ΦϨϯδͰද͞ΕΔ໘ੵɺ ͭ·Γ֬খ͘͞Ͱ͖Δɻ P
(|¯ xn µ | > " ) = P (|¯ xn µ | > " p n p n ) 5 2 " 2 n = 2 ¯ x = = 1 2 OΛ૿͍͚ͯ͠ ͜Ε̌ʹۙͮ͘ɻ νΣϏγΣϑͷෆࣜΛద༻ -
ຊൃදͷ༰ େͷ๏ଇ ਖ਼ن த৺ۃݶఆཧ ύϥϝʔλʔͷਪఆͱ ࠷๏ ۠ؒਪఆ ౷ܭతݕఆ هड़౷ܭͱਪଌ౷ܭ ֬ɾ֬ม
֬ ظ ଟ࣍ݩͷ֬ม ඪຊநग़ αϯϓϦϯά -
ਖ਼ن f(x) = 1 p 2⇡ 2 exp ✓ (x
µ) 2 2 2 ◆ f(x) = 1 p 2⇡ exp ✓ x 2 2 ◆ ਖ਼نͷີؔ ඪ४ਖ਼نͷີؔ 1 < x < 1 1 < x < 1 ౷ܭֶͷ͋ΒΏΔͱ͜ΖͰجຊతͳׂΛԋ͡Δ ͜ͷࣜͷΓཱͪΛߟ͑ͯΈ·͢ɻ -
̐ʣ͋ͱɺඍͨ࣌͠ʹܭࢉ͍͢͠Α͏ʹ ͱ͠·͢ɻͪΐͬͱԣʹ͕Γ·͢ɻ ਖ਼ن z = p 2y f ( z
) = exp ✓ 1 2 z2 ◆ -
ˡɹɹɹɹΑΓɹ ਖ਼ن ̑ʣ͜ͷG [ ͷ໘ੵ͕ͱͳΔΑ͏ʹʢ֬ͳͷͰɺͯ͢ ͷ͋Γ͑ΔࣄΛͨ͢ͱʹͳΔͷͰʣੵͯ͠ͱͳ ΔΑ͏ʹ͢Δඞཁ͕͋Γ·͢ɻ ˡΨεੵΑΓ Z 1
1 e y2 dy = p ⇡ Z 1 1 exp ✓ z2 2 ◆ dz = p 2 ⇡ Z 1 1 1 p 2 ⇡ exp ✓ z2 2 ◆ dz = 1 dz = p 2dy -
ਖ਼ن ̒ʣલϖʔδͰߦͬͨ໘ੵΛʹ͢ΔରԠΛߦ͏ͱඪ४ਖ਼ نͷີ͕ؔಘΒΕ·͢ɻ Z 1 1 1 p 2 ⇡
exp ✓ z2 2 ◆ dz -
ਖ਼ن ̓ʣมม z = x µ dz dx =
1 Λߦ͏ͱɺ ͱͳΔ͜ͱ͔Βɺ ໘ੵΛΩʔϓ͢ΔͨΊΛ͔͚Δਖ਼نͷີؔ ͕ಘΒΕΔɻ f(x) = Z 1 1 1 p 2⇡ 2 exp ✓ (x µ) 2 2 2 ◆ dx 1/ Ж -
ຊൃදͷ༰ େͷ๏ଇ ਖ਼ن த৺ۃݶఆཧ ύϥϝʔλʔͷਪఆͱ ࠷๏ ۠ؒਪఆ ౷ܭతݕఆ هड़౷ܭͱਪଌ౷ܭ ֬ɾ֬ม
֬ ظ ଟ࣍ݩͷ֬ม ඪຊநग़ αϯϓϦϯά -
ඪຊɹɹɹɹɹɹɹɹΛฏۉɹɺࢄɹΛͭ ҙͷ͔֬Βநग़͞Εͨͷͱ͢Δɻ͜ͷͱ͖ɺ த৺ۃݶఆཧ D = ( x1, · ·
· , xn) µ 2 ¯ x µ / p n , n ! 1 ʹର͢Δۃݶͱͯͭ͠·Γ N(0, 1) ඪ४ਖ਼نʹै͏֬มͰ͋Δ͜ͱ͕ࣔ͞ΕΔɻ αϯϓϧ ճ܁Γฦ͢ = 0.1, µ = 1 = 10, 2 = 1 2 = 100 ¯ x = p n = r 1 2n = r 1 0.01 ⇥ 10000 = r 1 100 = 1 10 Ξχϝʔγϣϯ63-IUUQTHJUIVCDPNNBUTVLFO2JJUB@$POUFOUTCMPCNBTUFS.BUI@DBGF@TUBUTBOJNFYQ@IJTUHJG -
ੵؔΛɹɹɹɹɹɹɹͱදݱ͢Δɻ த৺ۃݶఆཧ ಓ۩ͷ४උɿੵؔ g ( x ) = ext ͱஔ͍ͨͱ͖ͷظ
E[ ext] = Z 1 1 ext f ( x ) dx ΛUͷؔͱͯ͠Έͨ࣌ɺ͜ΕΛ֬มYͷੵؔ .PNFOU(FOFSBUJOH'VODUJPO ͱ͍͏ɻ M x (t) = E[ext] ੵؔͷॏཁͳੑ࣭ ֬มͷੵؔͱ֬มͷੵؔ ͕ͷۙͰҰக͢Δͱ͖ɺͦͷ̎ͭͷ֬ มಉ֬͡ʹै͏ɻ ূ໌ུ M x (t) My(t) x t = 0 y ˞ੵ͕ؔଘࡏ͢ΔͨΊʹͯ͢ͷ࣍Ͱੵ͕ଘࡏ͢Δඞཁ͕͋Δɻ -
த৺ۃݶఆཧ ·ͨɺɹΛςΠϥʔల։͢Δͱɺ g ( x ) = ext ext =
1 + xt + t 2 2!x 2 + · · · + tk k !xk + · · · Mx ( t ) = E[ ext] = E[1 + xt + t 2 2!x 2 + · · · + tk k !xk + · · · ] = 1 + t E[ x ] + t 2 2! E[ x 2] + · · · + tk k ! E[ xk] + · · · = 1 + xµ 0 1 + t 2 2!µ 0 2 + · · · + tk k !µ 0 k + · · · ͱͳΔͷͰɺੵؔ ݪ·ΘΓͷL࣍ͷੵ ͱɺల։Ͱ͖Δɻ -
த৺ۃݶఆཧ ੵه߸ԼͰͷඍՄೳੑΛԾఆͯ͠ɹɹɹΛLճඍ͢ Δͱ M x (t) d dtk Mx (
t ) = E[ xk ext] t = 0 d dtk Mx (0) = E[ xk] = µ 0 k ͱ͢Δͱɺ ͱͳΓɺݪ·ΘΓͷL࣍ͷੵ͕ಘΒΕΔ -
ͭ·Γɺɹɹɹɹɹͱ͢Δͱɺ த৺ۃݶఆཧ ਖ਼نͷੵؔ x ⇠ N ( µ, ) ͱ͢Δͱɺ
Mx(t) = E [ext ] = Z 1 1 ext 1 p 2⇡ 2 exp ✓ 1 2 (x µ) 2 2 ◆ dx z = x µ x = µ + z dx = dz Λར༻ͯ͠ -
͜͜Ͱɹɹɹɹɹͱ͓͍ͯɹɹɹɹΑΓ த৺ۃݶఆཧ M x( t ) = Z 1 1
e( µ + z ) t 1 p 2 ⇡ 2 exp ✓ 1 2 z2 ◆ dz = eµt Z 1 1 1 p 2 ⇡ exp ✓ tz 1 2 z2 ◆ dz = eµt Z 1 1 1 p 2 ⇡ exp ✓ 1 2 [ z2 2 tz 2t2 + 2t2 ] ◆ dz = eµt Z 1 1 1 p 2 ⇡ e 2t2 2 exp ✓ 1 2 ( z t ) 2 ◆ dz = eµte 2t2 2 Z 1 1 1 p 2 ⇡ exp ✓ 1 2 ( z t ) 2 ◆ dz z t = w dz = dw M x( t ) = eµte 2t2 2 Z 1 1 1 p 2 ⇡ exp ✓ w2 2 ◆ dw = eµt + 2t2 2 ͜Ε͕ਖ਼نͷ ੵؔ -
த৺ۃݶఆཧ ੵ͔ؔΒͷฏۉɺࢄΛٻΊΔ Ұ֊ඍ ೋ֊ඍ $IBJO3VMF (f · g)0 = f0
· g + f · g0 ੵͷඍ ( f g )0( x ) = f 0( g ( x )) g 0( x ) M0 x (t) = (µ + 2t)eµt + 2t2 2 M00 x (t) = (µ + 2t)2 ⇣ eµt + 2t2 2 ⌘ + 2 ⇣ eµt + 2t2 2 ⌘ = ⇣ eµt + 2t2 2 ⌘ {(µ + 2t)2 + 2} ·ͣɺੵؔͷҰ֊ɺೋ֊ඍΛٻΊΔɻ -
த৺ۃݶఆཧ Var[ x ] = E[ x 2] (E[ x
])2 = ( µ 2 + 2) ( µ )2 = 2 Var[ x ] = E[( x E[ x ])2] = E[ x 2 2E[ x ] x + E[ x ]2) = E[ x 2] 2E[ x ]2 + E[ x ]2 = E[ x 2] E[ x ]2 Α͘͏ࢄͷެࣜ ͜͜Ͱɹɹͱ͓͘ͱɺݪ·ΘΓͷ̍࣍ͷੵͱ࣍ͷ ੵ͕ಘΒΕΔͷͰɺ t = 0 E[ x ] = M 0 x (0) = ( µ + 2 · 0) eµ ·0+ 2·02 2 = µ E[ x 2] = M 00 x (0) = ⇣ eµ ·0+ 2·02 2 ⌘ {( µ + 2 · 0)2 + 2} = µ 2 + 2 -
·ͨɹɹɹɹɹɹɹʹର͢Δۃݶ த৺ۃݶఆཧ த৺ۃݶఆཧ ඪຊɹɹɹɹɹɹɹɹΛฏۉɹɺࢄɹΛͭҙͷ֬ ͔Βͷେ͖͞Oͷඪຊͱ͢Δɻඪຊ ͱ͓͘ͱɺ D = ( x1,
· · · , xn) µ 2 ¯ x µ / p n , n ! 1 ඪ४ਖ਼نɹɹɹɹͰ͋Δɻ N(0, 1) T = x1 + · · · + xn T nµ p n ࠓɹɹɹɹɹɹ͕ฏۉ̌ɺࢄɹͷ ਖ਼نʹै͏͜ͱΛূ໌͠·͢ɻ 2 T 0 = T nµ p n = ¯ x µ 1 / p n -
̍ʣҙͷ͔ΒOݸɺ֬มΛಘΔ ̎ʣOݸͷ֬มͷ͔ΒɹɹɹɹɹɹΛܭࢉ͠ɺ ɹɹͦͷੵؔΛٻΊΔ ̏ʣ̎ʣͰٻΊͨੵ͕ؔਖ਼نɹɹɹɹͷ ɹɹੵؔͱಉҰͰ͋Δ͜ͱΛࣔ͢ த৺ۃݶఆཧ ੵؔͷॏཁͳੑ࣭ ࠶ܝ ֬มͷੵؔͱ֬มͷੵؔ
͕ͷۙͰҰக͢Δͱ͖ɺͦͷ̎ͭͷ֬ มಉ֬͡ʹै͏ɻ M x (t) My(t) x t = 0 y த৺ۃݶఆཧͷূ໌ͷྲྀΕ T T0 = T nµ p n N(0, 2) ˞ੵ͕ؔଘࡏ͢Δ͜ͱΛ݅ͱ͢Δɻ -
த৺ۃݶఆཧ ূ໌ M xi (t) = 1 + µ0 1
t + µ0 2 t2 2! + µ0 3 t3 3! + · · · ݪ·ΘΓͷੵؔ ݪ·ΘΓͷ ࣍ͷੵ ݪ·ΘΓͷ ࣍ͷੵ ݪ·ΘΓͷ ࣍ͷੵ ฏۉ·ΘΓͷੵؔɺ M xi µ (t) = 1 + µ1t + µ2 t2 2! + µ3 t3 3! + · · · = 1 + 0 + 2 t2 2! + µ3 t3 3! + · · · ͷ͔̎ͭΒɺ -
த৺ۃݶఆཧ ΛͰׂͬͨɹɹɹͷੵؔɹɹɹ xi µ p n xi µ p n
M xi µ p n (t) = E[exi µ p n t] = 1 + 2 t2 2!n + µ3 t3 3!n3/2 + · · · + µk tk k!nk/2 + · · · = 1 + 2t2 2n + n 2n = 1 2n n n ! 0 n ! 0 ͱஔ͘ͱ ͷͱ͖ = 1 + 2t2 + n 2n -
த৺ۃݶఆཧ T 0 = x1 µ p n + x2
nµ p n + · · · + xn µ p n ͷੵؔΛߟ͑Δɻ ࣍ʹ = n X i=1 xi µ p n MT 0 (t) = MP n i =1 ⇣ xi µ p n ⌘(t) = E[e P n i =1 ⇣ xi µ p n ⌘ t ] = n Y i=0 E[e ⇣ xi µ p n ⌘ t ] = ✓ 1 + 1 n 2t2 + n 2 ◆n ಠཱͳͷͰ FͷఆٛΑΓ er ⌘ lim n!1 ⇣ 1 + r n ⌘n r r = lim n!1 ⇣ 1 + r n ⌘n -
த৺ۃݶఆཧ n ! 1 ͱ͢Ε lim n!1 MT 0 =
lim n!1 ✓ 1 + 1 n 2t2 + n 2 ◆n = e 2t2 2 S lim n!1 n = 0 ͷ ੵؔͱಉ͡ʂ Αͬͯ N(0, 2) T0 = T nµ p n ฏۉ̌ɺࢄɹͷਖ਼نʹ 2 ै͏͜ͱ͕Θ͔Δɻ -
ຊൃදͷ༰ େͷ๏ଇ ਖ਼ن த৺ۃݶఆཧ ύϥϝʔλʔͷਪఆͱ ࠷๏ ۠ؒਪఆ ౷ܭతݕఆ هड़౷ܭͱਪଌ౷ܭ ֬ɾ֬ม
֬ ظ ଟ࣍ݩͷ֬ม ඪຊநग़ αϯϓϦϯά -
ύϥϝʔλʔͷਪఆͱ࠷๏ ύϥϝʔλʔͷਪఆํ๏ʹɺਪఆɺ۠ؒਪఆͷ̎छ ྨ͕͋Δɻ ਪఆ ύϥϝʔλʔВΛ̍ͭͷɺͭ·Γ̍ ɹͰ ਪఆ͢Δਪఆํ๏ɻ ۠ؒਪఆ ύϥϝʔλʔВΛ۠ؒͰਪఆɺͭ·Γ
Ͱਪఆ͢Δํ๏ ˞ύϥϝʔλʔ͕֬తมಈΛ͢ΔΘ͚ ɹͰͳ͍͜ͱʹҙʢޙड़ʣ ✓0 = ˆ ✓(X1, · · · , Xn) ˆ ✓ lower (X 1 , · · · , Xn) 5 ✓ 0 5 ˆ ✓ upper (X 1 , · · · , Xn) -
ύϥϝʔλʔͷਪఆͱ࠷๏ ෆภੑ ਪఆྔͷظ͕ਅͷύϥϝʔλʔͱҰக͢Δɺ ͭ·Γਪఆྔͷظͱਅͷύϥϝʔλʔͱͷࠩ ͕̌Ͱ͋Δ͜ͱΛෆภੑͱ͍͏ Ұகੑ ඪຊOΛେ͖͍ͯ͘͘͠ͱɺਪఆྔ͕ਅͷύϥϝ ʔλʔʹ͖ۙͮɺOˠ㱣ͷ࣌ʹҰக͢Δੑ࣭ΛҰக ੑͱ͍͏ɻ ༗ޮੑ
ਪఆྔɹɹɹͷࢄ͕ɺͦͷཧతԼݶʹҰக͢ Δ߹༗ޮਪఆྔͱݺͿɻͭ·Γɺਪఆྔɹɹɹ ͷࢄ͕ΑΓখ͍͞ͷͷํ͕ΑΓ༗ޮɻ ޮͱݺͿɻ ਪఆʹز͔ͭੑ࣭͕͋Δɻ ˆ ✓(X) ˆ ✓(X) -
ύϥϝʔλʔͷਪఆͱ࠷๏ όΠΞεόϦΞϯεղ E[(ˆ ✓(X) ✓)2] = E[{(E[ˆ ✓(X)] ✓) +
(ˆ ✓(X) E[ˆ ✓(X)])}2] = E[(E[ˆ ✓(X)] ✓)2 + 2(E[ˆ ✓(X)] ✓)(ˆ ✓(X) E[ˆ ✓(X)]) + (ˆ ✓(X) E[ˆ ✓(X)])2] = (E[ˆ ✓(X)] ✓)2 + Var[ˆ ✓(X)] 㱺ظΛͱΔͱ̌ όΠΞε ෆภਪఆྔ όϦΞϯε ෆภਪఆྔόΠΞε͕̌ʹͳΔΑ͏ͳਪఆྔͷ͜ͱ Λࢦ͢ɻͦͷͨΊɺෆภਪఆྔͷฏۉೋޡࠩ E[ˆ ✓(X)] ✓ ฏۉೋޡࠩʮόΠΞεʯͱʮόϦΞϯεʯʹղͰ͖Δɻ E[(ˆ ✓(X) ✓)2] = Var[ˆ ✓(X)] ͱόϦΞϯεͷΈͱͳΔɻฏۉೋޡࠩΛ࠷খʹ͢Δෆภਪఆྔ ࢄΛ࠷খʹ͢ΔͷͱͳΔɻ -
ύϥϝʔλʔͷਪఆͱ࠷๏ E[¯ x ] = 1 n E[ T ]
= n · 1 n µ = µ ʮඪຊநग़ʯʹͯطʹݟͨΑ͏ʹظΛͱΔͱ ¯ x ͷΑ͏ʹɺύϥϝʔλʔͱ͘͠ͳΔͨΊɺฏۉʹؔ͢Δ ෆภਪఆྔͰ͋Δɻ ࣍ϖʔδͰࣔ͢Α͏ʹɺࢄͷෆภਪఆྔOͰׂΔͷͰͳ͘ OͰׂΔ s 2 = 1 n 1 n X i=1 ( xi ¯ x )2 ͱͳΔɻ ຊεϥΠυͷ࠷ॳͷ෦Ͱɺਪଌͷ߹OͰͳ͘OͰׂΔ ͱॻ͍ͨཧ༝͜ͷෆภੑΛ࣋ͨͤΔͨΊͰ͋ͬͨɻ -
ύϥϝʔλʔͷਪఆͱ࠷๏ Ұகਪఆྔ lim n!1 P {|¯ xn µ | =
" } = 0 ¯ xn ! µ in P ඪຊͷݸOݸΛ༻͍ͨਪఆྔ͕ͷ࣌ʹ ͷΑ͏ʹਅͷύϥϝʔλʔВʹ֬ऩଋ͢Δͱ͖ ਪఆྔɹɹɹҰகੑΛ࣋ͭͱ͍͏ɻ ҰகੑظͰͳ͘ɺਪఆྔͦͷͷ͕ਅͷύϥϝʔλʔͷ ʹۙͮ͘ͱ͍͏ੑ࣭Λදͨ͠ͷͰ͋Δɻ ˆ ✓n(X) n ! 1 ˆ ✓n(X) ! ✓ in P ˆ ✓n(X) ʮେͷ๏ଇʯͰࣔͨ͠ ͭ·Γ ɺɹ͕ฏۉɹʹର͢ΔҰகਪఆྔͰ͋Δ͜ͱΛද͍ͯͨ͠ɻ ¯ xn µ -
ύϥϝʔλʔͷਪఆͱ࠷๏ ༗ޮਪఆྔ ࠓճ֓ཁʹݶΔ όϦΞϯεɹɹɹɹɹΫϥϝϧϥΦͷෆࣜʹΑΓɺҰఆͷ ੍ݶ݅˞ ͷͱͰҙͷෆภਪఆྔɹɹɹʹର͠ɺ Var[ˆ ✓(X)] ύϥϝʔλʔВΛ࣋ͭ ֬ີؔ
ͱɺԼݶΛࣔ͢͜ͱ͕Ͱ͖Δɻ όϦΞϯε͢ͳΘͪਪఆྔͷࢄ͕͜ͷԼݶʹҰக͍ͯ͠Δͱ͖ɺ ͦͷਪఆྔΛ༗ޮਪఆྔͱ͍͍ɺ࠷ࢄ͕খ͍͞ਪఆྔͰ͋Δ͜ ͱ͕ࣔ͞ΕΔɻ ˆ ✓(X) Var[ˆ ✓ ( X )] = E "✓ @ @✓ log f ( X ; ✓ ) ◆2 # 1 = In( ✓ ) 1 ϑΟογϟʔใྔ ˞ ϑΟογϟʔใྔ͕ਖ਼Ͱ͋Δ͜ͱͱɺඍͱੵͷަ͕อো͞ΕΔͱ͍͏݅ͷͱɻ -
ύϥϝʔλʔͷਪఆͱ࠷๏ ࠷๏ʹΑΔύϥϝʔλʔͷਪఆ ඪຊɹɹɹɹɹɹɹɹɹ͕طʹಘΒΕ͍ͯΔͱͯ͠ɺɹີؔ ʹै͍ͬͯΔͱ͢Δɻ͜ͷඪຊͷಉ࣌֬ີؔ D = ( x1, · ·
· , xn) xi f ( xi) n Y i=1 f ( xi) n Y i=1 f ( xi | ✓ ) Ͱ͋ΔɻύϥϝʔλʔВΛ໌ࣔ͢Δͱ ͱͳΔɻ͜ͷ࣌ɺؔͷܗશ͘ಉ͡ͰɺมΛ Вͱ͠ɺɹΛݻఆͱͯ͠Έͨͱ͖ɺ͜ΕΛؔͱ͍͍ɺ ԼهͷΑ͏ʹද͢ɻ xi ` ( ✓ | x1, x2, · · · , xn) = n Y i=1 f ( xi | ✓ ) -
͍·ɺฏۉɺඪ४ภࠩͷਖ਼نʹै͏ཚݸ ɹɹɹ͕ಘΒΕɺͦͷ͕ ͩͬͨͱ͠Α͏ɻ ύϥϝʔλʔͷਪఆͱ࠷๏ <
> x1, x2, · · · , x10 σʔλͷ༷ࢠ ؔͷΠϝʔδ ಉ࣌֬ີؔԼهͷΑ͏ʹͳΔɻ f(x1, x2, · · · , x10 | µ, 2 ) = 10 Y i=1 1 p 2⇡ 2 exp ✓ 1 2 (xi µ) 2 2 ◆ -
ύϥϝʔλʔͷਪఆͱ࠷๏ ؔܗશ͘ಉ͕ͩ͡ɺؔԼهͷΑ͏ʹͳΔɻ `(µ, 2| x1, x2, · · · ,
x10) = 10 Y i=1 1 p 2⇡ 2 exp ✓ 1 2 (xi µ) 2 2 ◆ -
ύϥϝʔλʔͷਪఆͱ࠷๏ ࠷ਪఆ ࠷ਪఆͱɺؔͷ࠷େΛͱΔύϥϝʔλʔΛ୳͠ग़͢ ํ๏Ͱ͋ΓɺͦͷύϥϝʔλʔΛ࠷ਪఆྔͱ͍͍ ✓ ⇤ = arg max ✓
`(✓ | x1, x2, · · · , xn) ͱද͢ɻ ܭࢉΛߦ͏্ͰରΛͱͬͨ log `(✓ | x1, · · · , xn) ⌘ L(✓ | x1, · · · , xn) Λ͏ํ͕ɺੵΛʹ͢Δ͜ͱ͕Ͱ͖ͯศརͳ͜ͱ͕ଟ͍ɻ MPH୯ௐ૿ՃؔͳͷͰେখ͕ؔมΘΒͣɺ-Λ࠷େʹ͢ΔВ ͕ಉ࣌ʹɹΛ࠷େʹ͢ΔВͱͳΔɻ ` -
ύϥϝʔλʔͷਪఆͱ࠷๏ ࠷ਪఆྔΛղੳతʹٻΊΔɿਖ਼نͷྫ ରΛͱͬͯ ɹɹʹؔͯ͠࠷େΛٻΊΔͨΊɺภඍ͢Δͱ࠷ਪఆྔ͕ ٻΊΒΕΔɻ µ, 2 L(µ, 2| x1,
x2, · · · , x10) = n 2 (2⇡) n 2 log 2 1 2 2 n X i=1 (xi µ) 2 @L @µ = 1 2 2 n X i=1 ( xi µ )2 ) n X i=1 xi = nµ ) µ ⇤ = 1 n n X i=1 xi `(µ, 2| x1, x2, · · · , xn) = n Y i=1 1 p 2⇡ 2 exp ✓ 1 2 (xi µ) 2 2 ◆ -
ύϥϝʔλʔͷਪఆͱ࠷๏ @L @ 2 = n 2 1 2 +
1 2( 2)2 n X i=1 ( xi µ )2 = 0 ) 1 2( 2)2 n X i=1 ( xi µ )2 = n 2 2 ) 2⇤ = 1 n n X i=1 ( xi µ )2 ࠷ਪఆͰಘΒΕΔෆภਪఆྔʹͳΒͳ͍͜ͱʹҙɻ ʢࢄΛաখਪఆͯ͠͠·͏ɻʣ 2⇤ -
ຊൃදͷ༰ େͷ๏ଇ ਖ਼ن த৺ۃݶఆཧ ύϥϝʔλʔͷਪఆͱ ࠷๏ ۠ؒਪఆ ౷ܭతݕఆ هड़౷ܭͱਪଌ౷ܭ ֬ɾ֬ม
֬ ظ ଟ࣍ݩͷ֬ม ඪຊநग़ αϯϓϦϯά -
Vඪ४ਖ਼نʹै͏֬มɹɹɹɹɹɹɺWࣗ༝Nͷ ΧΠೋʹै͏֬มͱ͠ɺͦΕͧΕಠཱͱ ͨ͠ͱ͖ɺ֬มU ۠ؒਪఆ U u ⇠ N(0, 1) t
= u p v/m v ⇠ 2(m) ɺ۠ؒ 㱣 㱣 ʹ͓͍ͯUʹै͍ɺͦͷ֬ີؔ f(t) = m+1 2 p m⇡ m 2 ✓ t2 m + 1 ◆ m+1 2 -
पลΛͱΔͱɺ ɹɹɹɹɹɹɺɹɹɹɹɹɺɺɹɹʹର͠ ಉ࣌֬ີؔ ۠ؒਪఆ U DPOU ɿུূ u ⇠ N(0,
1) v ⇠ 2(m) v > 0 1 < u < +1 f ( u, v ) = 1 p 2 ⇡ exp ✓ u2 2 ◆ (1 / 2) n/2 ( n/ 2) vn/2 1e v/2 ͱͳΔɻ͜͜Ͱมม ΧΠೋͷີؔ ඪ४ਖ਼نͷີؔ t = u p v/m x = v ɺɹɹɹɹΛ͓͜ͳ͍ࣜมܗͷͷͪUʹؔ͢Δ f(t) = m+1 2 p m⇡ m 2 ✓ t2 m + 1 ◆ m+1 2 ɹ͕ಘΒΕΔɻ ˞͖ͪΜͱͨ͠ূ໌ʮֶγϦʔζཧ౷ܭֶ Ҵ֞એੜ ʯɹQςΟʔࢀর ΨϯϚؔ (z) = Z 1 0 tz 1e t dt -
ඪຊɹɹɹɹɹɹɹɹɹΛɹɹɹɹɹɹɹͱ͢Δͱɺ ඪຊฏۉɹɹɹɹɹɹɹɹɹͱͳΔɻ ۠ؒਪఆ ͜ͷUΛԿʹ͏͔ͱ͍͏ͱɺฏۉɺࢄ͕ະͷਖ਼ن ͔Βͷ֬ඪຊʹରͯ͠ɺฏۉɹͷ۠ؒਪఆʹ͑Δɻ µ D = ( x1,
· · · , xn) xi ⇠ N ( µ, 2) ¯ x ⇠ N ( µ, 2 /n ) ¯ x ˞ূ໌ʮֶγϦʔζཧ౷ܭֶ Ҵ֞એੜ ʯɹQඪຊฏۉͱඪຊࢄͷࢀর 1 2 n X i=1 ( xi ¯ x )2 ⇠ 2 n 1 ·ͨɺඪຊฏۉͱͷภࠩͷೋΛࢄͰׂͬͨͷɺ ࣗ༝OͷΧΠೋʹै͏ɻ -
ෆภࢄΛɹɹɹɹɹɹɹɹɹɹɹɹͱ͓͘ɻ ɺ ۠ؒਪఆ u = ¯ x µ / p
n ⇠ N (0 , 1) v = 1 2 n X i=1 ( xi ¯ x )2 ⇠ 2 n 1 ͱ͢Δͱɺ t = u p v/ ( n 1) = ¯ x µ / p n · " 1 2 1 ( n 1) n X i=1 ( xi ¯ x )2 # 1/2 = ¯ x µ 1 / p n · 1 p s 2 = ¯ x µ s/ p n ⇠ tn 1 s 2 = 1 n 1 n X i=1 ( xi ¯ x )2 ෆภࢄ s2 ˢࢄМ͕ফ͑ͨʂ -
ɹɹɹΛ৴པͱΑͼɺͷ͕۠ؒઃఆͰ͖Δɻ Uͷ྆֎ଆͷ֬Λ͠߹ΘͤͨͷΛɹͱͨ͠ͱ͖ɺ ۠ؒਪఆ P ✓ tn 1;↵/2 5 ¯ x
µ s/ p n 5 tn 1;↵/2 ◆ = 1 ↵ tn 1;↵/2 tn 1;↵/2 ໘ੵ ↵/2 ໘ੵ ↵/2 ໘ੵ 1 ↵ 1 ↵ 1 ↵ P ✓ ¯ x tn 1;↵/2 s p n 5 µ 5 ¯ x + tn 1;↵/2 s p n ◆ = 1 ↵ [ tn 1;↵/2 , tn 1;↵/2 ] Uʹै͏ ࣜมܗΛ͓͜ͳ͏ͱɺ ͜Εɺूஂͷฏۉ ͕͜ͷ۠ؒʹೖΔ֬ ɹɹɹͰ͋Δ͜ͱΛࣔ͠ ͍ͯΔɻ µ 1 ↵ -
ɹɹɹΛ৴པͱΑͼɺͷ͕۠ؒઃఆͰ͖Δɻ Uͷ྆֎ଆͷ֬Λ͠߹ΘͤͨͷΛɹͱͨ͠ͱ͖ɺ ۠ؒਪఆ P ✓ tn 1;↵/2 5 ¯ x
µ s/ p n 5 tn 1;↵/2 ◆ = 1 ↵ tn 1;↵/2 tn 1;↵/2 ໘ੵ ↵/2 ໘ੵ ↵/2 ໘ੵ 1 ↵ 1 ↵ 1 ↵ P ✓ ¯ x tn 1;↵/2 s p n 5 µ 5 ¯ x + tn 1;↵/2 s p n ◆ = 1 ↵ [ tn 1;↵/2 , tn 1;↵/2 ] Uʹै͏ ࣜมܗΛ͓͜ͳ͏ͱɺ ͜Εɺूஂͷฏۉ ͕͜ͷ۠ؒʹೖΔ֬ ɹɹɹͰ͋Δ͜ͱΛࣔ͠ ͍ͯΔɻ µ 1 ↵ ͨͩ͠ɺฏۉЖ͕֬తมಈΛͯ͠ɺ͜ͷ ۠ؒʹೖΔ֬Ͱͳ͘ɺ܁Γฦ͠ඪຊΛ αϯϓϦϯά͠ɺಉ༷ͷ۠ؒΛઃఆ͢Δͱɺ ͜ͷઃఆ͕ͨ۠ؒ͠ฏۉЖΛଊ͑Δ͕֬ ЋͰ͋Δɺͱ͍͏͜ͱʹҙ -
ຊൃදͷ༰ େͷ๏ଇ ਖ਼ن த৺ۃݶఆཧ ύϥϝʔλʔͷਪఆͱ ࠷๏ ۠ؒਪఆ ౷ܭతݕఆ هड़౷ܭͱਪଌ౷ܭ ֬ɾ֬ม
֬ ظ ଟ࣍ݩͷ֬ม ඪຊநग़ αϯϓϦϯά -
౷ܭతݕఆ Λ୯७Խͨ͠ͷΛࠓճར༻͢Δ͜ͱͱ͠ɺࢄطͰɹɹɹ ͱ͢Δɻ͜ͷͱ͖ɺฏۉɹɹɹͰ͋Δ͜ͱͷݕఆΛߦ͏ɻ ؼແԾઆɿ ରཱԾઆɿ ྆ଆݕఆ = 1 µ =
0 H0 : µ0 = 0 H1 : µ 6= µ0 ݕఆͷઃఆ ΑͬͯɺूஂͷؼແԾઆͰԼهͷΑ͏ͳ͕Ծఆ͞ ΕΔɻ -
౷ܭతݕఆ ਅͷঢ়ଶ έʔε̍ɿؼແԾઆ͕ਖ਼͍͠ έʔε̎ɿରཱԾઆ͕ਖ਼͍͠ ݕ ఆ ݁ Ռ ؼແԾઆ ؼແԾઆͷʹै͍ɺ
࠾ҬʹೖΔ ରཱԾઆͷʹै͍ɺ࠾ ҬʹೖΔ ରཱԾઆ ؼແԾઆͷʹै͍ɺ غ٫ҬʹೖΔ ରཱԾઆͷʹै͍ɺغ ٫ҬʹೖΔ ͯ͞ɺ͜͏ͬͯઃఆͨ͠غ٫Ҭʹରͯ͠ɺਅͷԾઆʢؼແԾઆɺରཱ Ծઆʣͱɺݕఆ݁ՌʢؼແԾઆΛड͚ೖΕΔɺରཱԾઆΛड͚ೖΕΔʣ ͷºͷͭͷύλʔϯ͕͋Γ͏Δɻදʹද͢ͱԼهͷ௨Γɻ -
ࢀߟ ɾՊܥͷͨΊͷ౷ܭ֓ ΨοτϚϯɺΟϧΫε ɹIUUQTXXXBNB[PODPKQEQ ɾཧ౷ܭֶσʔλղੳͷํ๏ ܒ ɹIUUQTXXXBNB[PODPKQEQ ɾݱཧ౷ܭֶจࣾݱܦࡁֶબॻ
ଜজ௨ ɹIUUQTXXXBNB[PODPKQEQ ɾֶγϦʔζཧ౷ܭֶ Ҵ֞એੜ ɹIUUQTXXXBNB[PODPKQEQ -
*% UPUBM@CJMM UJQ HFOEFS TNPLFS EBZ UJNF TJ[F
'FNBMF /P 4VO %JOOFS .BMF /P 4VO %JOOFS .BMF /P 4VO %JOOFS .BMF /P 4VO %JOOFS 'FNBMF /P 4VO %JOOFS .BMF /P 4VO %JOOFS .BMF /P 4VO %JOOFS .BMF /P 4VO %JOOFS … … … … … … … … ૬ؔ ར༻σʔλɿҿ৯ళͷސ٬ผࢧֹ͍ۚσʔλ ग़యɿIUUQTHJUIVCDPNNXBTLPNTFBCPSOEBUBCMPCNBTUFSUJQTDTW -
ࣜͰද͢ͱɾɾɾ r = 1 n Pn i=1 ( xi ¯
x )( yi ¯ y ) q 1 n Pn i=1 ( xi ¯ x )2 q 1 n Pn i=1 ( yi ¯ y )2 ૬ؔͱ -
ࣜͰද͢ͱɾɾɾ ͖ͬ͞ͷඪ४ภࠩͱҰॹʂ r = 1 n Pn i=1 ( xi
¯ x )( yi ¯ y ) q 1 n Pn i=1 ( xi ¯ x )2 q 1 n Pn i=1 ( yi ¯ y )2 ૬ؔͱ -
ࣜͰද͢ͱɾɾɾ ͖ͬ͞ͷඪ४ภࠩͱҰॹʂ r = 1 n Pn i=1 ( xi
¯ x )( yi ¯ y ) q 1 n Pn i=1 ( xi ¯ x )2 q 1 n Pn i=1 ( yi ¯ y )2 ͍ͯ͠ΔͷͰඞͣϓϥε ૬ؔͱ -
ࣜͰද͢ͱɾɾɾ r = 1 n Pn i=1 ( xi ¯
x )( yi ¯ y ) q 1 n Pn i=1 ( xi ¯ x )2 q 1 n Pn i=1 ( yi ¯ y )2 ͜ΕΛڞࢄͱ͍͏ ૬ؔͱ -
ڞࢄ 1 n n X i=1 ( xi ¯ x
)( yi ¯ y ) ԣ࣠ͷฏۉ͔Βͷڑ ภࠩ ฏۉΑΓখ͍͞ͱ͜Ζ ϚΠφεʹͳΔɻ ૬ؔͱ -
ڞࢄ 1 n n X i=1 ( xi ¯ x
)( yi ¯ y ) ॎ࣠ͷฏۉ͔Βͷڑ ภࠩ ฏۉΑΓখ͍͞ͱ͜Ζ ϚΠφεʹͳΔɻ ૬ؔͱ -
ฏۉ ૬ؔ 1 n n X i=1 ( xi ¯
x )( yi ¯ y ) ૬ؔͱ ڞࢄ -
ϚΠφε ϚΠφε ϓϥε ϓϥε ૬ؔ ૬ؔͱ 1 n n X
i=1 ( xi ¯ x )( yi ¯ y ) ڞࢄ -
ϚΠφε ϚΠφε ϓϥε ϓϥε ૬ؔ ૬ؔͱ 1 n n X
i=1 ( xi ¯ x )( yi ¯ y ) ڞࢄ -
ϚΠφε ϚΠφε ϓϥε ϓϥε ૬ؔ ૬ؔͱ 1 n n X
i=1 ( xi ¯ x )( yi ¯ y ) ڞࢄ -
ϚΠφε ϚΠφε ϓϥε ϓϥε ૬ؔ ૬ؔͱ 1 n n X
i=1 ( xi ¯ x )( yi ¯ y ) ڞࢄ