コンテンツの人気度を考慮したN次創作活動のモデル化
全文
(2) Vol.2016-DBS-163 No.10 Vol.2016-IFAT-123 No.10 2016/9/13. ใॲཧֶձڀݚใࠂ IPSJ SIG Technical Report. ڞ༗͓ΑͼɼطଘϞσϧΛ༻͍ͨੜϞσϧͷ੍࡞ΛՄೳ ͱ͢Δ web αΠτΛެ։ͨ͠ɽ൴Β 10.4%ͷϞσϧ͕ ଞͷϞσϧͷੜ͍ͯͬͳʹݩΔ͜ͱɼੜ࡞׆ಈ ࠷Ͱୈ̐ੈ·Ͱଓ͍ͨ͜ͱΛใࠂ͍ͯ͠ΔɽCheliotic Β [6] ΦϯϥΠϯԻָίϛϡχςΟͷ ccMixter*3 ʹ͓͚ ΔԻָίϯςϯπͷ N ࣍࡞׆ಈͷௐࠪΛߦͬͨɽௐࠪͷ ݁ՌɼN ࣍࡞׆ಈʹΑͬͯίϛϡχςΟશମͷίϯςϯ πྔ͕ඈ༂తʹ૿Ճ͢Δͱʹڞɼίϯςϯπͷଟ༷ੑߴ ͘ͳΔ͜ͱ͕໌Β͔ʹͳͬͨɽHamasaki Β [1] ಈըڞ ༗αʔϏεͷχίχίಈը *4 ্Ͱͷੜ࡞׆ಈΛର ͱ͠ɼΦϦδφϧίϯςϯπͱੜίϯςϯπͷؒͷࣗ໌ ͳࢠؔΛѻ͍ɼ֤ΦϦδφϧίϯςϯπͷੜίϯς ϯπͳͲͷ౷ྔܭΛੳͨ͠ɽҎ্ͷͰڀݚੜίϯ ςϯπͷ੍࡞͕ʮͲͷΑ͏ʹʯߦΘΕͨΛੳ͍ͯ͠Δͷ ʹରͯ͠ɼզʑʮͳͥʯੜίϯςϯπ੍͕࡞͞Ε͔ͨ ʹண͠ɼͦͷཁҼΛਪఆ͢ΔͨΊͷϞσϧΛఏҊ͢Δɽ. [ࣾձ׆ಈʹ͓͚ΔӨڹͷϞσϧԽ] ࣾձ׆ಈͰͷϢʔβ ؒͷӨ͕ڹਪఆͰ͖ΔͱɼӨྗڹͷେ͖͍Ϣʔβͷಛఆ [7]. 3. Ϟσϧ ࣾձ׆ಈΛϞσϧԽ͢ΔࡍɼΞΠςϜʹର͢ΔϢʔβ ͷΈʢຊͰڀݚΦϦδφϧίϯςϯπͷັྗʹରԠʣ͓ ΑͼϢʔβؒͷӨڹΛߟྀ͢Δͷ͕ҰൠతͰ͋Δ [4], [11]ɽ ͔͠͠ɼN ࣍࡞׆ಈʹ͓͍ͯɼΫϦΤʔλؒͷӨ͕ڹ ੜίϯςϯπͷ੍࡞ΛҾ͖͢͜ىཁҼʹͳ͍ͬͯΔ͜ͱ ใࠂ͞Ε͓ͯΒͣ [1], [5], [6]ɼͦͷଘࡏٙΘ͍͠ɽ ͦͷΘΓʹɼʮrich-get-richerʯʹݱͮ͘ج͕ଘࡏ͢Δ Ͱ͋Ζ͏͜ͱ͕͜Ε·Ͱͷใࠂ͔Βߟ͑ΒΕΔ [1]ɽͦ͜ ͰզʑɼΦϦδφϧίϯςϯπ͓Αͼੜίϯςϯπͷ ਓ ͕ؾN ࣍࡞׆ಈʹ͓͍ͯॏཁͳׂΛՌͨ͢ͱԾ ఆ͢ΔɽຊষͰɼ̐ͭͷཁҼʢΦϦδφϧίϯςϯπͷ ັྗɼϢʔβؒͷӨڹɼΦϦδφϧίϯςϯπͷਓؾɼ ੜίϯςϯπͷਓؾʣΛؚΊͨϞσϧʹ͍ͭͯड़Δ͕ɼ զʑ͕ఏҊ͢Δͷ͔ͦ͜ΒϢʔβؒͷӨڹΛআ͍ͨ̏ͭ ͷཁҼΛؚΊͨϞσϧͰ͋Δɽ. 3.1 ߸هͷఆٛ. ͷਪન [8] ʹ͓͍ͯ༗༻Ͱ͋ΔͨΊɼଟ͘ͷਪఆख. ʮҜࢠͷ 3D ϞσϧʯʮطଘָۂΛՎͬͨಈըʯͳͲͷυ. ๏͕ఏҊ͞Ε͖ͯͨɽओཁͳΞϓϩʔνͷͻͱͭɼಠཱ. ϝΠϯ͓Αͼσʔλͷ؍ଌ ؒظT ͕༩͑ΒΕͨͱ͖ɼI Λ. ΧεέʔυϞσϧ [9] ͳͲͷใ֦ࢄϞσϧΛ༻͍Δ͜ͱ. ͋Δ web αʔϏεʹ࣌ࠁ 0 ͔Β T ͷؒʹߘ͞Εͨରυ. Ͱ͋Δɽ͜ͷϞσϧͰࢄ࣌ؒΛԾఆ͍ͯ͠Δ͕ɼSaito. ϝΠϯͷΦϦδφϧίϯςϯπͷू߹ͱ͢Δɽ(tpij , upij ). Β [10] ࿈ଓ࣌ؒΛѻ͏ͨΊʹϙΞιϯաఔʹͮ͘جϞ. ΛΦϦδφϧίϯςϯπ i ∈ I ͷ j ൪ʹߘ͞Εͨੜ. σϧΛఏҊͨ͠ɽ൴ΒͷϞσϧͰɼϢʔβؒͷӨڹͷ༗. ίϯςϯπͱ͢Δɽ۩ମతʹɼΫϦΤʔλ͕ upij ∈ U ͕. ແΛද͢άϥϑใ͕ඞཁͰ͋Δ͕ɼIwata Β [4] άϥ. ࣌ࠁ tpij ʹ i ͷੜίϯςϯπΛߘͨ͜͠ͱΛද͢ɽU . ϑใ͕༩͑ΒΕͳͯ͘ϢʔβؒͷӨؔڹΛਪఆ͢Δ. શΫϦΤʔλू߹Ͱ͋ΔɽҰൠੑΛࣦ͏͜ͱͳ͘ɼੜί. ϞσϧɼShared Cascade Poisson ProcessʢSCPPʣΛఏҊ. ϯςϯπͷߘΠϕϯτߘ࣌ࠁʹؔͯ͠ঢॱʹฒΒ. ͨ͠ɽTanaka Β [11] SCPP Λ֦ு͠ɼͷߪങΛҾ. Ε͍ͯΔͱ͢Δʢj < j ′ ʹରͯ͠ tpij ≤ tpij ′ ʣɽJi Λ؍ଌظ. ͖ͨ͜͠ىཁҼΛਪఆ͢ΔϞσϧΛఏҊͨ͠ɽ൴ΒϢʔ. ؒதʹߘ͞Εͨ i ͷશੜίϯςϯπͱ͢Δͱɼi ͷ. βؒͷӨʹڹՃ͑ͯɼϢʔβ͕ϝσΟΞࠂ͔Βड͚ΔӨ. i Ͱ ੜίϯςϯπͷߘΠϕϯτू߹ Di = {(tpij , upij )}Jj=1. ڹߟྀ͠ɼSCPP ͷߟ͕͑ͷߪങߦಈͷϞσϧԽʹ. ද͞ΕΔɽैͬͯɼશͯͷΦϦδφϧίϯςϯπͷੜί. ༗༻Ͱ͋Δ͜ͱΛࣔͨ͠ɽզʑͷϞσϧ SCPP ͓Αͼ. ϯςϯπͷߘΠϕϯτू߹ D = {Di }i∈I ͱද͞ΕΔɽ. Tanaka Β [11] ͷϞσϧΛ֦ுͨ͠ͷͰ͋Δ͕ɼ࣍ͷ. ΫϦΤʔλ web αʔϏε্ͰΦϦδφϧίϯςϯπͷ. ͰطଘϞσϧͱҟͳΔɿ ʢ̍ʣطଘϞσϧͰϢʔβʹΑͬ. o ) ਓͮ͘جʹؾϥϯΩϯάΛӾཡͰ͖Δͱ͢Δɽ(toik , rik. ͯબ͞ΕͨΞΠςϜʢߪೖ͞ΕͨͳͲʣͷӨڹߟ. Λ i ∈ I ͷ k ൪ͷϥϯΩϯάೖΓΠϕϯτͱ͠ɼi ͕࣌. ྀ͢Δඞཁ͕ͳ͔ͬͨɽ͔͠͠ɼN ࣍࡞׆ಈʹ͓͍ͯɼ. o ҐʹϥϯΩϯάೖΓͨ͜͠ͱΛද͢ɽҰൠੑ ࠁ toik ʹ rik. ΫϦΤʔλʹΑ੍ͬͯ࡞͞Εͨੜίϯςϯπࣗମଞͷ. Λࣦ͏͜ͱͳ͘ɼϥϯΩϯάೖΓΠϕϯτϥϯΩϯάೖ. ΫϦΤʔλͷ࡞׆ಈʹӨڹΛ༩͑ΔɽͦͷͨΊզʑΦ. Γͨ࣌͠ࠁʹؔͯ͠ঢॱʹฒΒΕ͍ͯΔͱ͢ΔɽKio Λ. Ϧδφϧίϯςϯπͱੜίϯςϯπͷ྆ํͷӨڹΛѻ͑. i ͕؍ଌ͍͓ͯʹؒظϥϯΩϯάೖΓͨ͠ճͱ͢Δͱɼ. ΔΑ͏ʹ SCPP Λ֦ு͢Δɽʢ̎ʣطଘϞσϧͰΞΠς. o i i ͷϥϯΩϯάೖΓΠϕϯτͷू߹ Oi = {(toik , rik )}k=1. Ϝͷਓ͕ؾϢʔβͷΞΠςϜબʹ༩͑ΔӨڹҰఆͰ. ͱද͞ΕɼશΦϦδφϧίϯςϯπͷϥϯΩϯάೖΓΠϕ. ͋ΔͱԾఆ͍͕ͯͨ͠ɼզʑͦͷӨڹ࣌ؒʹΑͬͯม. ϯτू߹ O = {Oi }i∈I ͱද͞ΕΔɽ. Ko. Խ͢ΔͱԾఆ͢ΔɽզʑͷఏҊϞσϧͰίϯςϯπͷਓ. ΫϦΤʔλੜίϯςϯπͷਓͮ͘جʹؾϥϯΩϯ. ؾϥϯΩϯάͱΫϦΤʔλͷϥϯΩϯάӾཡߦಈΛߟྀ. άӾཡͰ͖Δͱ͢ΔɽΦϦδφϧίϯςϯπͷ߹ͱಉ. ͢Δ͜ͱͰɼΦϦδφϧίϯςϯπ͓Αͼੜίϯςϯπ. c ༷ɼ(tcik , rik ) Λ i ͷੜίϯςϯπͷ k ൪ͷϥϯΩϯάೖ. ͷ࣌ؒʹԠͨ͡ਓؾΛѻ͏ɽ. ΓΠϕϯτͱ͢ΔɽKic Λ i ͷੜίϯςϯπ͕؍ଌʹؒظ. *3. ͓͍ͯϥϯΩϯάೖΓͨ͠ճͱ͢Δͱɼi ͷੜίϯςϯ. *4. http://ccmixter.org http://www.nicovideo.jp. ⓒ 2016 Information Processing Society of Japan. 2.
(3) Vol.2016-DBS-163 No.10 Vol.2016-IFAT-123 No.10 2016/9/13. ใॲཧֶձڀݚใࠂ IPSJ SIG Technical Report (c) 䜸䝸䝆䝘䝹䝁䞁䝔䞁䝒 䛾ேẼ. (e) 䠎␒┠䛾ὴ⏕䝁䞁䝔䞁䝒䛾ேẼ. 0. T. Rate. Rate. Rate. (a) 䜸䝸䝆䝘䝹䝁䞁䝔䞁䝒 䛾㨩ຊ. 0. T. T. T. Rate. (f) 䛶䛾せᅉ䜢⪃៖䛧䛯䝺䞊䝖. Rate. Rate 0. ਤ 1. 0. (d) 䠍␒┠䛾ὴ⏕䝁䞁䝔䞁䝒䛾ேẼ. (b) 䜽䝸䜶䞊䝍䛾ᙳ㡪. 0. T. 0. T. ΫϦΤʔλ u ͕࣌ࠁ t ʹΦϦδφϧίϯςϯπ i ͷੜίϯςϯπΛߘ͢ΔϨʔτɽ Kc. ′. c i )}k=1 ɼ πͷϥϯΩϯάೖΓΠϕϯτͷू߹ Ci = {(tcik , rik. u ͷӨͳͱྗڹΔɽe−γp (t−t ) ɼύϥϝʔλ γp ≥ 0 ͷ. શΦϦδφϧίϯςϯπͷશੜίϯςϯπͷϥϯΩϯά. ͱͰ࣌ؒͷܦաͱʹڞӨݮ͕ྗڹਰ͢Δ͜ͱΛද͢ɽ ਤ 1ʢbʣͰɼ̎ਓͷΫϦΤʔλ͕ i ͷੜίϯςϯπ. ೖΓΠϕϯτू߹ C = {Ci }i∈I ͱද͞ΕΔɽ. Λߘ͍ͯ͠Δɽ৭Ͱද͞Εͨ̍ਓͷΫϦΤʔλΛ u′. 3.2 ཁҼ. ͱ͢Δͱɼu′ ͷӨྗڹ u′ ͕ੜίϯςϯπΛߘͨ͠. 3.2.1 ΦϦδφϧίϯςϯπͷັྗ ΦϦδφϧίϯςϯπ i ͷਓ͕ؾ͍߹ͰɼΫϦΤʔ λ u ͕ i Λັྗతͩͱ͡ײΕɼu i ͷੜίϯςϯπ Λ੍࡞͢Δ͔͠Εͳ͍ɽi ͷັྗ i ͷ༷ʑͳಛྔ͔ ΒҾ͖͜͞ىΕ͏Δɽͨͱָ͑ۂͷ߹ɼϝϩσΟՎ ࢺͳͲͷಛྔ͕ߟ͑ΒΕΔɽΦϦδφϧίϯςϯπͷັ ྗ͔Βड͚ΔӨڹͷେ͖͞ΫϦΤʔλʹΑͬͯҟͳΓɼ ·ͨͦͷӨ͍ͯͮجʹڹੜίϯςϯπΛߘ͢ΔϨʔτ ࣌ࠁ 0 ͔Β T ͷؒͰҰఆͰ͋ΔͱԾఆ͢Δʢਤ 1ʢaʣ ʣ ɽ ͜͜Ͱɼ࣌ࠁ t ʹ͓͚ΔϨʔτͱɼΫϦΤʔλ͕࣌ࠁ t ʹ i ͷੜίϯςϯπΛߘ͢Δ֬Λද͢ɽ͜ͷԾఆʹ ͖ͮجɼi ͷັྗ͕ཁҼͱͳͬͯΫϦΤʔλ u ͕ i ͷੜ ίϯςϯπΛߘ͢ΔϨʔτΛ࣍ࣜͰٻΊΔɽ. u ͕ΦϦδφϧίϯςϯπͷັྗʹӨڹΛड͚Δ߹ u∈U. 3.2.3 ΦϦδφϧίϯςϯπͷਓؾ ΦϦδφϧίϯςϯπ i ͕ফඅऀͷؒͰਓ͋ͰؾΕɼi ΛͨݟΫϦΤʔλ u ͕ i ͷੜίϯςϯπΛ੍࡞͢Δ͔ ͠Εͳ͍ɽ3.1 અͰड़ͨΑ͏ʹɼຊͰڀݚΫϦΤʔλ ͕ΦϦδφϧίϯςϯπͷਓؾϥϯΩϯάΛӾཡͰ͖Δ ͱԾఆ͢Δɽ̎ͭͷίϯςϯπ͕ϥϯΩϯάೖΓ͍ͯ͠Δ ͱ͖ɼϥϯΩϯά্Ґͷίϯςϯπͷํ͕ߴ͍ӨྗڹΛ࣋ ͭͱ͍͏ԾઆΛཱͯΔɽ·ͨɼ3.2.2 ߲ͱಉ༷ʹɼΦϦδφ ϧίϯςϯπͷਓ͕ؾ༩͑ΔӨڹΫϦΤʔλʹΑͬͯҟ ͳΓɼӨྗڹ࣌ؒͷܦաͱݮʹڞਰ͢ΔͱԾఆ͢Δɽ͜ ίϯςϯπΛߘ͢ΔϨʔτΛ࣍ࣜͰٻΊΔɽ. αi ≥ 0 ΦϦδφϧίϯςϯπ i ͷັྗΛද͢ɽθ0u ≥ 0. P. ܦաͱʹڞӨྗڹݮগ͍ͯ͘͠ɽ. ͷԾఆʹ͖ͮجɼi ͷਓؾͷӨ Ͱڹu ͕࣌ࠁ t ʹ i ͷੜ. fi (u) = αi θ0u .. ͍Ͱ͋Γɼ. ࣌ࠁ tpi1 Ͱ αu′ θu′ u ʢਤதͷ h1 ʹ૬ʣͰ͋Γɼ࣌ؒͷ. θ0u = 1 Λຬͨ͢ɽਤ 1ʢaʣͰɼ੨͍. rb(r′ )ω θ e−γo (t−t′ ) i −1u ho(i,t′ ,r′ ) (t, u) = 0. t′ < t. if. otherwise.. όʔͷߴ͕͞ αi θ0u ʹ૬͢Δɽ. r′ i ͷ࣌ࠁ t′ ʹ͓͚ΔϥϯΩϯάͰͷॱҐΛɼؔ rb. 3.2.2 ΫϦΤʔλͷӨڹ. ॱҐʹΑΔόΠΞεΛද͢ɽϢʔβͷ web ݁ࡧݕՌͷӾ. ΫϦΤʔλ u′ ͕ i ͷੜίϯςϯπΛߘ͢ΔͱɼΫϦ. ཡ࣌ͷৼΔ͍Λੳͨ͠Ͱڀݚɼweb ϖʔδͷॱҐ͕. Τʔλ u u ͔ΒӨڹΛड͚ͯ i ͷੜίϯςϯπΛ੍. ͘ͳΔͱʹڞɼӾཡ͞ΕΔ֬ʹܹٸԼ͕Δ͜ͱ͕ใࠂ. ࡞͢Δ͔͠Εͳ͍ɽ͜ͷͱ͖ɼͨͱ͑ u ͕ u′ ͷϑΝ. ͞Ε͍ͯΔ [3]ɽ͜ͷӾཡϞσϧʹ͖ͮجɼຊͰڀݚॱҐ. ϯͰ͋Εɼu′ u ʹରͯ͠େ͖ͳӨྗڹΛ࣋ͭΑ͏ʹɼ. ʹΑΔόΠΞεΛ rb(r ′ ) =. ′. 1 r′. ʹΑΓٻΊΔɽωi ≥ 0 i ͷ. u ͕༩͑ΔӨڹͷେ͖͞ΫϦΤʔλʹΑͬͯҟͳΔͱ. ਓʹؾΑΔӨྗڹΛɼθ−1u ≥ 0 u ͕ΦϦδφϧίϯςϯ. Ծఆ͢Δɽ·ͨɼϢʔβؒͷใ֦ࢄաఔͷϞσϧԽʹ฿. πͷਓ͔ؾΒӨڹΛड͚Δ߹͍Λද͠ɼ. ͍ [12]ɼΫϦΤʔλͷӨྗڹ࣌ؒͷܦաͱݮʹڞਰ͢Δ. Λຬͨ͢ɽe. ͱԾఆ͢Δɽ͜ΕΒͷԾఆʹ͖ͮجɼ࣌ࠁ t′ ʹ i ͷੜί. ͷܦաͱʹڞӨݮ͕ྗڹਰ͢Δ͜ͱΛද͢ɽ. ′. ′. ϯςϯπΛߘͨ͠ u ͷӨͰڹɼu ͕࣌ࠁ t ʹ i ͷੜί. otherwise. ′. αu′ ≥ 0 u ͷӨྗڹΛɼθu′ u ≥ 0 u ͕ u ͔ΒӨڹΛड P ͚Δ߹͍Λද͠ɼ u∈U \u′ θu′ u = 1 Λຬͨ͢ɽU \ u′ ′. U ͔Β u Λআ͍ͨू߹Λද͢ɽͭ·Γɼα θ ⓒ 2016 Information Processing Society of Japan. u′. θ−1u = 1. ɼύϥϝʔλ γo ≥ 0 ͷͱͰ࣌ؒ. ਤ 1ʢcʣͰɼi ϥϯΩϯάʹ̐ճग़͍ͯ͠ݱΔɽ࠷. toi1 ʹ͓͚ΔӨྗڹ rb(r′ )ωi θ−1u ʢਤதͷ h2 ʹ૬ʣͰ. t′ < t. if. ′. u∈U. ॳͷϥϯΩϯάೖΓ࣌ͷॱҐΛ r′ ɼ࣌ࠁΛ toi1 ͱ͢Δͱɼ. ϯςϯπΛߘ͢ΔϨʔτΛ࣍ࣜͰٻΊΔɽ. α ′ θ ′ e−γp (t−t′ ) u uu g(i,t′ ,u′ ) (t, u) = 0. −γo (t−t′ ). P. u′ u. ′. u ͔Β. ͋Γɼ࣌ؒͷܦաͱʹڞӨྗڹݮগ͍ͯ͘͠ɽ. 3.2.4 ੜίϯςϯπͷਓؾ ΫϦΤʔλ u′ ੍͕࡞ͨ͠ i ͷੜίϯςϯπ͕ফඅऀͷ ؒͰਓ͋ͰؾΕɼͦͷੜίϯςϯπΛͨݟΫϦΤʔλ. u ɼͨͱ͑ u′ ͷϑΝϯͰͳ͍ͱͯ͠ɼi ͷੜίϯ 3.
(4) Vol.2016-DBS-163 No.10 Vol.2016-IFAT-123 No.10 2016/9/13. ใॲཧֶձڀݚใࠂ IPSJ SIG Technical Report. ςϯπΛ੍࡞͢Δ͔͠Εͳ͍ɽ3.2.3 ߲Ͱड़ͨԾઆ͓ ′. ΑͼԾఆͱಉ༷ʹɼi ͷੜίϯςϯπ͕࣌ࠁ t ʹ r Ґʹ ϥϯΩϯάೖΓͨ͠߹ɼͦͷӨ Ͱڹu ͕࣌ࠁ t ʹ i ͷ ੜίϯςϯπΛߘ͢ΔϨʔτΛ࣍ࣜͰٻΊΔɽ. rb(r′ )σ θ e−γd (t−t′ ) i −2u hd(i,t′ ,r′ ) (t, u) = 0. if. ද 1. t′ < t. otherwise.. σʔληοτͷ౷ྔܭ. |I|. |O|. |D|. |C|. |U |. ՎͬͯΈͨ. 4,035. 64,973. 199,320. 67,627. 18,715. གྷͬͯΈͨ. 396. 30,925. 9,420. 22,954. 1,153. ԋͯ͠Έͨ. 583. 38,726. 5,526. 20,492. 692. ′. ίϯςϯπͷັྗɼΦϦδφϧίϯςϯπͷਓؾɼੜί ϯςϯπͷਓؾɼͷ̏ͭͷཁҼΛߟྀ͢Δ͜ͱ N ࣍࡞ ׆ಈΛϞσϧԽ͢Δࡍʹ༗ޮ͔ʢ4.2 અʣ ɽN ࣍࡞׆ಈΛ. σi ≥ 0 i ͷੜίϯςϯπͷਓʹؾΑΔӨྗڹΛɼ. ϞσϧԽ͢ΔࡍʹɼίϯςϯπͷਓؾϥϯΩϯάͷॱҐ. θ−2u ≥ 0 u ͕ੜίϯςϯπͷਓ͔ؾΒӨڹΛड͚Δ P ′ ߹͍Λද͠ɼ u∈U θ−2u = 1 Λຬͨ͢ɽe−γd (t−t ) ɼ. ʹରͯ͠ͲͷΑ͏ͳόΠΞεΛ͔͚Δͷ͕༗ޮ͔ʢ4.3 અʣ ɽ. γd ≥ 0 ͷͱͰ࣌ؒͷܦաʹΑΔӨྗڹͷݮਰΛද͢ɽ. ਤ 1ʢdʣ͓ΑͼʢeʣͰɼi ͷ̍൪ͷੜίϯςϯ. πͱ̎൪ͷੜίϯςϯπͷӨྗڹΛͦΕͧΕද͍ͯ͠ Δɽਤ 1ʢdʣͷ࠷ॳͷϥϯΩϯάೖΓͷॱҐΛ r′ ͱ͢Δ ͱɼ࠷ॳʹϥϯΩϯάೖΓͨ࣌͠ࠁ tci1 ʹ͓͚ΔӨྗڹ. rb(r′ )σi θ−2uʢਤதͷ h3 ʹ૬ʣͰ͋Γɼ࣌ؒͷܦաͱڞ ʹӨྗڹݮগ͍ͯ͘͠ɽ. 4.1 σʔληοτ ຊ࣮Ͱݧɼಈըڞ༗αʔϏεʮχίχίಈըʯͷ N ࣍ ࡞׆ಈσʔλΛ༻ͨ͠ɽχίχίಈըͰɼಛʹԻָ ίϯςϯπͷ N ࣍࡞׆ಈ͕ΜͰ͋Γɼ2016 7 ݄ͷ ࣌Ͱ 14 ສ݅Ҏ্ͷΦϦδφϧίϯςϯπͱ 60 ສ݅Ҏ্ ͷੜίϯςϯπ͕ߘ͞Ε͍ͯΔɽେ෦ͷΦϦδφϧ ίϯςϯπ VOCALOID ͱݺΕΔՎ߹ٕज़Λ༻͍ ੍ͯ࡞͞Εָͨ͋ͰۂΔɽੜίϯςϯπʹؔͯ͠ɼΦ ϦδφϧۂΛՎ͏ʮՎͬͯΈͨʯ ɼΦϦδφϧ߹ʹۂΘͤͯ. 3.3 ੜίϯςϯπͷߘϨʔτ 3.2.1 ߲͔Β 3.2.4 ߲Ͱड़ͨཁҼʹ͖ͮجɼΫϦΤʔλ. གྷΔʮགྷͬͯΈͨʯɼΦϦδφϧۂΛԋ͢Δʮԋͯ͠. u ͕࣌ࠁ t ʹΦϦδφϧίϯςϯπ i ͷੜίϯςϯπΛ. Έͨʯͷ̏ͭͷυϝΠϯΛରͱ࣮ͯ͠ݧΛߦͬͨɽຊ࣮. ߘ͢ΔϨʔτΛ࣍ࣜͰද͢ɽ. Ͱݧ 2010 1 ݄ 1 ͔Β 2013 3 ݄ 31 ͷؒʹߘ. λi (t, u) = fi (u) +. X. g(i,t′ ,u′ ) (t, u). (t′ ,u′ )∈Dit\u. +. X. (t′ ,r ′ )∈O. ho(i,t′ ,r′ ) (t, u) +. X. hd(i,t′ ,r′ ) (t, u). (t′ ,r ′ )∈C. it. it. ͜͜ͰɼDit\u = {(t′ , u′ )|(t′ , u′ ) ∈ Di and t′ < t ∧ u′ 6=. u}ɼOit = {(t′ , r′ )|(t′ , r′ ) ∈ Oi and t′ < t}ɼCit = {(t′ , r′ )|(t′ , r′ ) ∈ Ci and t′ < t} Ͱ͋Δɽλi (t, u) ਤ 1 ʢfʣதͷ h4 ʹ૬͢Δɽ ؍ଌσʔλ DɼOɼC ͕༩͑ΒΕͨͱ͖ɼD ͷؔ ࣍ࣜͰද͞ΕΔɽ. Β 2013 6 ݄ 30 ͷؒʹߘ͞ΕͨੜίϯςϯπΛର ͱͨ͠ɽ2010 1 ݄ 1 ͔Β 2013 3 ݄ 31 ͷؒͷ σʔλΛֶशσʔλɼ2013 4 ݄ 1 ͔Β 2013 6 ݄ 30 ͷؒͷσʔλΛςετσʔλͱͨ͠ɽ͍ͣΕͷυϝΠϯ Ͱɼֶश͍͓ͯʹؒظੜίϯςϯπ͕̎݅ະຬͷΦϦ δφϧίϯςϯπ͓Αͼɼੜίϯςϯπͷߘ͕݅̏ ݅ະຬͷΫϦΤʔλσʔληοτ͔Βআ͍ͨɽ χίχίಈըͰɼΦϦδφϧίϯςϯπ͓Αͼ্هͷ ̏υϝΠϯͷੜίϯςϯπͦΕͧΕʹ͍ͭͯɼ୯ҐͰ ਓؾίϯςϯπͷ্Ґ 100 ݅ΛݟΔ͜ͱ͕Ͱ͖Δɽ̍. P (D|O, C, α, ω, σ, Θ, γ) ! J Z TX i Y Y = exp − λi (t, u)dt λi (tpij , tpij ). i∈I. ͞ΕͨΦϦδφϧίϯςϯπ͓Αͼɼ2010 1 ݄ 1 ͔. 0. u∈U. j=1. ύϥϝʔλ α = {αl }l∈I∪U ɼω = {ωi }i∈I ɼσ = {σi }i∈I ɼ. Θ = {θ u }u∈U+ ɼθ u = {θuu′ }u′ ∈U \u ɼγ = {γp , γo , γd } Ͱ ͋ΔɽU+ U ∪ {0, −1, −2} Λද͠ɼ0ɼ−1ɼ−2 ͦΕͧ ΕɼΦϦδφϧίϯςϯπͷັྗɼΦϦδφϧίϯςϯπ ͷਓؾɼੜίϯςϯπͷਓؾɼʹ૬͢ΔԾΫϦΤʔ λΛද͢ɽIwata Β [4] ʹ฿͍ɼϕΠζਪఆʹ֤͍ͯͮج ύϥϝʔλͷࣄલΛԾఆ͠ɼ֬త EM ΞϧΰϦζϜ Λ༻͍Δ͜ͱͰύϥϝʔλ͓ΑͼࣄલͷϋΠύʔύϥ ϝʔλΛਪఆ͢Δɽࢴ໘ͷ߹্ɼৄࡉׂѪ͢Δɽ. 4. ఆྔతධՁ ຊষͰɼ࣮ݧΛ௨ͯ࣍͠ͷٙʹ͑ΔɽΦϦδφϧ. ⓒ 2016 Information Processing Society of Japan. ͷϥϯΩϯάɼલͷӾཡίϝϯτͳͲͷ͍ͭ͘ ͔ͷࢦඪΛ࡞ʹݩ͞ΕΔɽզʑ 2010 1 ݄ 1 ͔Β. 2013 6 ݄ 30 ͷؒͷɼΦϦδφϧίϯςϯπ͓Αͼ̏ υϝΠϯͷੜίϯςϯπͷ্Ґ 100 ݅ͷϥϯΩϯάσʔ λΛऩूͨ͠ɽϥϯΩϯάσʔλ͕୯ҐͰ͋ΔͨΊɼ ੜίϯςϯπͷߘ࣌ࠁ୯ҐͰѻͬͨɽ ද 1 ʹຊ࣮ͨ͠༻Ͱݧσʔληοτͷ౷ྔܭΛࣔ͢ɽ. 4.2 ཁҼͷ߹͕ͤ༩͑ΔӨڹ [ൺֱϞσϧ] OattɼUinfɼOpopɼDpop ΛͦΕͧΕΦ ϦδφϧίϯςϯπͷັྗɼΫϦΤʔλͷӨڹɼΦϦδ φϧίϯςϯπͷਓؾɼੜίϯςϯπͷਓ͢ͱؾΔɽ3 ষͰड़ͨΑ͏ʹɼզʑ OattɼOpopɼDpop ΛؚΉϞ σϧ͕࠷ޮՌతͰ͋Δͱ͍͏ԾઆΛཱͯͨɽ͜ͷԾઆ Λ͢ূݕΔͨΊʹɼ࣍ͷ̒छྨͷϞσϧΛൺֱͨ͠ɿʢ1ʣ. Oattɼ ʢ2ʣOatt+Uinfɼ ʢ3ʣOatt+Uinf+Opop+Dpopɼ ʢ4ʣ 4.
(5) Vol.2016-DBS-163 No.10 Vol.2016-IFAT-123 No.10 2016/9/13. ใॲཧֶձڀݚใࠂ IPSJ SIG Technical Report. Linear. Uniform 㻝㻢. 㻝㻡 㻝㻞 㻥 㻢 㻟 㻜. 㻤. ḷ䛳䛶䜏䛯. *. 㻢. *. 䠍䞄᭶. ਤ2. 㻜. 䠏䞄᭶. 䠍䞄᭶. 㻟 㻞. *. 㻞. 䠎䞄᭶. * *. 㻠. *. 㻠. ㋀䛳䛶䜏䛯. 㻝. 䠎䞄᭶. 㻜. 䠏䞄᭶. ₇ዌ䛧䛶䜏䛯 * *. *. 䠍䞄᭶. 䠏䞄᭶. 䠎䞄᭶. 㻝㻞. ḷ䛳䛶䜏䛯. 䠍䞄᭶. ਤ 3. 㻢. ㋀䛳䛶䜏䛯. 䠎䞄᭶. 䠏䞄᭶. 䠍䞄᭶. 㻟. ₇ዌ䛧䛶䜏䛯 *. 㻝. 䠎䞄᭶. 䠏䞄᭶. 㻜. * *. 㻞. *. 㻞 㻜. * *. 㻠. *. 㻠. Ϟσϧ͝ͱͷෛͷରɽॎ࣠ɿෛͷରɽԣ࣠ɿςε. * *. 㻤 㻜. Reciprocal 㻠. 㻤. 䠍䞄᭶. 䠎䞄᭶. 䠏䞄᭶. ϥϯΩϯάதͷॱҐʹΑΔόΠΞεͷ͔͚͔ͨͷख๏͝ͱͷ ෛͷରɽ. τؒظɽ. Oatt+Opopɼ ʢ5ʣOatt+Dpopɼ ʢ6ʣOatt+Opop+Dpopɽ. ද 2 ਪఆ͞ΕͨཁҼͷൺʢ%ʣ ՎͬͯΈͨ གྷͬͯΈͨ ԋͯ͠Έͨ. ͨͱ͑ Oatt+Uinf ɼOatt ͱ Uinf ΛؚΉϞσϧͰ͋Δ. ཁҼ. ͜ͱΛද͢ɽ̒छྨͷϞσϧͷதͰɼʢ2ʣ CCPP [4] ʹ. Oatt. 14.6. 17.3. 42.5. ରԠ͠ɼʢ6ʣ͕զʑͷఏҊϞσϧʹରԠ͢Δɽ. Opop. 40.0. 21.7. 40.0. Dpop. 45.4. 61.0. 17.5. [ධՁࢦඪ] ఏҊϞσϧͷଥੑΛධՁ͢ΔͨΊʹɼϞσ ϧͷ༧ଌਫ਼Λςετσʔλʹର͢ΔෛͷରʹΑͬ. ͯ͢ಉ͡ӨྗڹΛ࣋ͭͱԾఆ͍ͯ͠Δɽ֤ख๏ͷධՁʹ. ͯධՁ͢ΔɽςετؒظΛ T ͔Β T ′ ͱͨ͠ͱ͖ɼςετ. ɼ4.2 અͱಉ༷ʹෛͷରΛ༻͍ͨɽ. ؒظதͷੜίϯςϯπͷߘΠϕϯτ (t, u) ͷෛͷର. Πϯͷશͯͷςετ͍͓ͯʹؒظଞͷ̎ख๏ͷ༧ଌਫ਼Λ. ࣍ࣜͰද͞ΕΔɽ. L=−. X i∈I. −. Z. T. T. ′. X. u∈U. [݁Ռ] ਤ 3 ʹ݁ՌΛࣔ͢ɽReciprocal ख๏͕શͯͷυϝ. λi (t, u)dt. !. X. ্ճ͍ͬͯΔ͜ͱ͔ΒɼΫϦΤʔλͷϥϯΩϯάӾཡ࣌ͷ. logλi (t, u).. (t,u)∈Ditest. Ditest i ͷςετσʔλͰ͋ΓɼL ͷ͕খ͍͞΄Ͳ༧ଌ ਫ਼͕ߴ͍͜ͱΛද͢ɽςετؒظͷ͞ͷӨڹௐΔ ͨΊɼςετؒظΛ̍ϲ݄ʢ2013 4 ݄ 1 ͔Β 2013 . 4 ݄ 30 ʣ͔Β̏ϲ݄ʢ2013 4 ݄ 1 ͔Β 2013 6 ݄ 30 ʣ·Ͱ̍ϲ݄୯ҐͰ૿Ճͤͨ͞ɽ [݁Ռ] ݁ՌΛਤ 2 ʹࣔ͢ɽ֤ςετ࠷Ͱؒظ༧ଌਫ਼. ৼΔ͍Λߟྀͨ͠ Reciprocal ख๏͕ N ࣍࡞׆ಈΛϞ σϧԽ͢Δ্Ͱ࠷༗༻Ͱ͋Δͱ͑ݴΔɽ. 5. ఆੑతධՁ ఏҊϞσϧΛ༻͍ͯɼֶश͚͓ʹؒظΔ֤ੜίϯςϯ πͷߘ͕̏ͭͷ֤ཁҼ͔ΒӨڹΛड͚ͨ߹͍Λਪఆ͢ Δ͜ͱͰɼఆੑతͳධՁΛߦͬͨɽ. 5.1 υϝΠϯ͕࣋ͭಛੑ. ͷߴ͔ͬͨϞσϧΛʮ*ʯͰࣔ͢ɽ ʮՎͬͯΈͨʯ͓Αͼ. ςετؒظதͷ̏ͭͷ֤ཁҼͷӨڹͷେ͖͞ͷൺΛ. ʮགྷͬͯΈͨʯͰɼఏҊϞσϧͷ Oatt+Opop+Dpop ͕. ද 2 ʹυϝΠϯ͝ͱʹࣔ͢ɽ̏ͭͷཁҼͷൺυϝΠϯ. શͯͷςετ࠷Ͱؒظ༧ଌਫ਼͕ߴ͔ͬͨɽʮԋͯ͠. ʹΑͬͯେ͖͘ҟͳ͍ͬͯͨɽʮՎͬͯΈͨʯͰɼOpop. ΈͨʯͰɼఏҊϞσϧ͕࠷ߴ͍ਫ਼ʹͳΒͳ͔ͬͨ. ͓Αͼ Dpop ͷൺ͕ߴ͘ɼOatt ͷൺ͕͍ɽ͜ͷ݁. ͕ɼશͯͷυϝΠϯͷશͯͷςετ҆Ͱؒظఆͯ͠ߴ͍ਫ਼. Ռ͔ΒɼʮՎͬͯΈͨʯͷΫϦΤʔλྲྀߦʹහͰײɼί. Λه͍ͯͨ͠ɽʮԋͯ͠ΈͨʯͰ࠷ߴ͍ਫ਼Λه. ϯςϯπͷਓؾΛॏࢹͯ͠ੜίϯςϯπΛ੍࡞͍ͯ͠Δ. ͨ͠ϞσϧɼʮՎͬͯΈͨʯʮགྷͬͯΈͨʯʹ͓͚. ͜ͱ͕༧͞ΕΔɽʮགྷͬͯΈͨʯΧςΰϦͰ Dpop ͷ. Δ༧ଌਫ਼͕͍͜ͱ͋Γɼ҆ఆੑʹ͚ܽΔ݁ՌͰ͋ͬ. ൺ͕ߴ͍ɽʮགྷͬͯΈͨʯͰɼશͯͷΫϦΤʔλ͕ಠ. ͨɽ͜ΕΒͷ݁Ռ͔Βɼൺֱͨ̒͠छྨͷϞσϧͷதͰɼ. ࣗͷৼΛߟ͑ΒΕΔΘ͚Ͱͳ͍ɽͦͷͨΊɼ͋ΔΫϦ. OattɼOpopɼDpop Λߟྀͨ͠ఏҊϞσϧ͕ N ࣍࡞׆. Τʔλ͕ৼΛߟ͑ͯགྷͬͨੜίϯςϯπΛߘ͠ɼͦ. ಈΛϞσϧԽ͢Δ্Ͱ࠷༗༻Ͱ͋Δͱ͑ݴΔɽ. ͷޙɼଞͷΫϦΤʔλ͕ͦͷৼΛਅࣅͯੜίϯςϯπ. 4.3 ॱҐʹΑΔόΠΞεͷ͔͚͔ͨͷൺֱ [ηοςΟϯά] 3.2.3 ߲Ͱड़ͨΑ͏ʹɼఏҊϞσϧͰ ਓؾϥϯΩϯάதͷॱҐͷٯΛ༻͍ͯॱҐʹΑΔό ΠΞεΛ͔͚ΔʢҎ߱ɼ͜ͷख๏Λ Reciprocal ͱͿݺʣɽ ఏҊख๏ͷ༗༻ੑΛ͢ূݕΔͨΊʹɼ࣍ͷ̎ͭͷख๏ͱͷ ൺֱΛߦ͏ɽ̍ͭͷख๏ʢLinear ख๏ʣɼॱҐʹΑΔ. Λߘ͢Δ͜ͱ͕Α͘ߦΘΕ͍ͯΔɽදͷ݁ՌɼఏҊϞ σϧʹΑͬͯ͜ͷΑ͏ͳಛੑΛଊ͑ΒΕ͍ͯΔ͜ͱΛࣔ͠ ͍ͯΔɽʮԋͯ͠ΈͨʯͰɼOatt ͷൺ͕ߴ͍ͨΊɼΫ ϦΤʔλྲྀߦʹͱΒΘΕͣʹɼࣗͷ͖ͳָۂΛԋ ͯ͠ੜίϯςϯπΛ੍࡞͍ͯ͠Δ͜ͱ͕༧͞ΕΔɽ. 5.2 ཁҼͷ࣌ؒతਪҠ. o 101−rik c 100 ɽrb(rik ). ਤ 4 ʹɼυϝΠϯ͝ͱʹ͋ΔΦϦδφϧίϯςϯπͷ. ಉ༷ʹ͢ࢉܭΔɽ͜ͷख๏ͰɼίϯςϯπͷॱҐ͕Լ. ੜίϯςϯπͷ੍࡞ΛҾ͖ͨ͜͠ىཁҼͷ࣌ؒతਪҠͷ. ͕ͬͯɼͦͷӨྗڹ Reciprocal ख๏΄ͲݮʹܹٸΒ. ਪఆ݁ՌΛࣔ͢ɽԣ࣠̍ϲ݄୯ҐͰද͞هΕ͓ͯΓɼ࠷. ͳ͍ͱԾఆ͍ͯ͠Δɽ̎ͭͷख๏ʢUniform ख๏ʣɼ. ॳͷ݄ੜίϯςϯπ͕ॳΊͯߘ͞Ε݄ͨͰ͋Δɽॎ. ॱҐʹΑΔόΠΞεΛߟྀ͠ͳ͍ɽͭ·ΓɼॱҐʹΑΒͣ. ֤݄࣠ͷੜίϯςϯπͷ૯ߘΛද͢ɽάϥϑͷ. = 1 ͱ͠ɼϥϯΩϯάͷίϯςϯπ. ੨ɼΦϨϯδɼͦΕͧΕ OattɼOpopɼDpop ʹΑͬͯ. o όΠΞεΛઢʹܗมԽͤ͞Δɿrb(rik ). o rb(rik ). =. c rb(rik ). ⓒ 2016 Information Processing Society of Japan. =. 5.
(6) Vol.2016-DBS-163 No.10 Vol.2016-IFAT-123 No.10 2016/9/13. ใॲཧֶձڀݚใࠂ IPSJ SIG Technical Report 䜸䝸䝆䝘䝹䝁䞁䝔䞁䝒䛾㨩ຊ䠄Oa 䠅 120. 25. ḷ䛳䛶䜏䛯. 100. 䜸䝸䝆䝘䝹䝁䞁䝔䞁䝒䛾ேẼ䠄Opop䠅. ὴ⏕䝁䞁䝔䞁䝒䛾ேẼ䠄Dpop䠅 25. ㋀䛳䛶䜏䛯. 20. 20. 15. 15. 10. 10. 5. 5. ₇ዌ䛧䛶䜏䛯. 80 60 40 20 0 Apr-10. Oct-10. Apr-11. Oct-11. Apr-12. Oct-12. 0 Nov-10. 0 May-11. Nov-11. May-12. Nov-12. Jul-10. Jan-11. Jul-11. Jan-12. Jul-12. Jan-13. ਤ 4 ͋ΔΦϦδφϧίϯςϯπʹணͨ͠ͱ͖ͷɼ̏ͭͷ֤ཁҼʹΑͬͯҾ͖͜͞ىΕͨͱ ਪఆ͞Εͨੜίϯςϯπͷߘɽॎ֤݄࣠ͷੜίϯςϯπͷ૯ߘΛද͢ɽ. 6. ·ͱΊ ຊߘͰੜίϯςϯπͷ੍࡞ΛҾ͖ͨ͜͠ىཁҼΛ ਪఆ͢ΔͨΊͷϞσϧΛఏҊͨ͠ɽࠓޙͷ՝ͱͯ͠ɼ. Thingiverse ͳͲͷଞυϝΠϯͷ N ࣍࡞׆ಈσʔλʹఏ ҊϞσϧΛద༻ͯ͠༗༻ੑΛ͢ূݕΔ͜ͱɼ5 ষͰࣔͨ͠ Α͏ͳੳ݁ՌΛ web ্ͰӾཡ͢ΔͨΊͷΠϯλϑΣʔε ΛఏҊ͢Δ͜ͱͳͲ͕͋͛ΒΕΔɽ. ँࣙ ਤ 5 ʮགྷͬͯΈͨʯΧςΰϦͰͷੜͷաఔɽ. Ҿ͖͜͞ىΕͨͱਪఆ͞ΕͨੜίϯςϯπͷߘΛද ͢ɽ͜͜ͰɼυϝΠϯ͝ͱͷಛ͕ݱΕ͍ͯΔ͜ͱ͕Θ ͔ΔɽʮՎͬͯΈͨʯͰɼॳظͷஈ֊Ͱ Oatt ͱ Opop. ຊڀݚͷҰ෦ɼจ෦ՊֶলՊֶڀݚඅิॿۚ׆ڀݚಈε λʔτࢧԉʢ՝൪߸ 15H06887ʣ͓ΑͼՊֶٕज़ৼߏػڵ OngaCREST ϓϩδΣΫτͷࢧԉΛड͚ͨɽ. ࢀߟจݙ [1]. ͷӨ͕ڹେ͖͘ɼͦͷ ޙDpop ͷӨ͕ڹେ͖͘ͳΔ͜ͱ͔ Βɼॳʹظߘ͞Εͨੜίϯςϯπ͕ਓͳʹؾΓɼͦͷ ޙDpop Λॏࢹ͢ΔΫϦΤʔλ͜ͷΦϦδφϧίϯςϯ. [2]. πͷੜίϯςϯπΛߘͨ͜͠ͱ͕༧͞ΕΔɽʮགྷͬ ͯΈͨʯͰɼॳظͷஈ֊ͰಠࣗʹৼΛߟ͑ΒΕΔݶ. [3]. ΒΕͨΫϦΤʔλ͕ੜίϯςϯπΛߘ͠ʢ࠷ॳͷ̎ϲ ݄ͷ੨෦ʹ૬ʣ ɼͦͷޙɼͦΕΒͷੜίϯςϯπʹӨ. [4]. ڹΛड͚ͨଟ͘ͷΫϦΤʔλ͕৽͍͠ੜίϯςϯπΛ ߘ͍ͯ͠Δ͜ͱ͕ಡΈऔΕΔɽʮԋͯ͠ΈͨʯͰɼؒظ. [5]. ͷલͰ Opop Dpop Λॏࢹ͢ΔΫϦΤʔλʹΑΔ ੜίϯςϯπͷߘ͕ΜͰ͋Δ͕ɼޙؒظʹίϯςϯ πͷਓ͕ؾਰ͑ͯɼOatt Λॏࢹ͢ΔΫϦΤʔλʹΑͬͯ ੜίϯςϯπ͕ߘ͞Εଓ͚͍ͯΔ͜ͱ͕Θ͔Δɽ. [6] [7]. 5.3 ΦϦδφϧίϯςϯπΛ͢ͱىΔੜͷաఔ ਤ 5 ʹʮགྷͬͯΈͨʯͷ͋ΔΦϦδφϧίϯςϯπͷ. [8]. ੜίϯςϯπͷੜաఔΛࣔ͢ɽਤதͰ 0 ΦϦδφϧί. [9]. ϯςϯπΛɼj ≥ 1 j ൪ʹߘ͞Εͨੜίϯςϯπ Λද͢ɽࣈؒͷࢬԼͷࣈͷίϯςϯπ্͕ͷࣈͷ. [10]. ίϯςϯπͷӨڹΛड੍͚ͯ࡞͞Εͨ͜ͱΛද͢ɽ͜ͷྫ Ͱɼ9 ൪ͱ 11 ൪ͷੜίϯςϯπ͕࣍ͷੈͷଟ. [11]. ͘ͷੜίϯςϯπͷ੍࡞ΛҾ͖͓ͯ͜͠ىΓɼN ࣍࡞ ׆ಈʹ͓͍ͯॏཁͳׂΛՌ͍ͨͯ͠Δ͜ͱ͕Θ͔Δɽ· ͨɼ͜ͷྫͰୈ 10 ੈ·Ͱੜίϯςϯπʢ30 ൪ɼ. 32 ൪ɼ33 ൪ʣ͕ଘࡏ͢Δ͜ͱ͕Θ͔Δɽ ⓒ 2016 Information Processing Society of Japan. [12]. M. Hamasaki et al.: “Network analysis of massively collaborative creation of multimedia contents: Case study of hatsune miku videos on nico nico douga”, UXTV, pp. 165–168 (2008). M. Goto: “Grand challenges in music information research”, Dagstuhl Follow-Ups: Multimodal Music Processing, 3, pp. 217–225 (2012). T. Joachims et al.: “Accurately interpreting clickthrough data as implicit feedback”, SIGIR, pp. 154–161 (2005). T. Iwata et al.: “Discovering latent influence in online social activities via shared cascade poisson processes”, KDD, pp. 266–274 (2013). K. Eto et al.: “Modulobe: A creation and sharing platform for articulated models with complex motion”, ACE, pp. 305–308 (2008). G. Cheliotis and J. Yew: “An analysis of the social structure of remix culture”, C&T, pp. 165–174 (2009). X. Song et al.: “Information flow modeling based on diffusion rate for prediction and ranking”, WWW, pp. 191–200 (2007). X. Song et al.: “Personalized recommendation driven by information flow”, SIGIR, pp. 509–516 (2006). J. Yang and J. Leskovec: “Modeling information diffusion in implicit networks”, ICDM, pp. 599–608 (2010). K. Saito et al.: “Learning continuous-time information diffusion model for social behavioral data analysis”, ACML, pp. 322–337 (2009). Y. Tanaka et al.: “Inferring latent triggers of purchases with consideration of social effects and media advertisements”, WSDM, pp. 543–552 (2016). S. Myers and J. Leskovec: “On the convexity of latent social network inference”, Advances in Neural Information Processing Systems 23, pp. 1741–1749 (2010).. 6.
(7)
関連したドキュメント
18)Kobayashi S, Takeda T, Enomoto M, Tamori A, Kawada N, Habu D, et al.: Development of hepatocellular carci- noma in patients with chronic hepatitis C who had a sus- tained
et al.: Selective screening for coronary artery disease in patients undergoing elective repair of abdominal
Consistent with previous re- ports that Cdk5 is required for radial migration of cortical neurons in mice (Gilmore et al., 1998; Ohshima et al., 2007), radial migration of
Cichon.M,et al.1997, Social Protection and Pension Systems in Central and Eastern Europe, ILO-CEETCentral and Eastern European TeamReport No.21.. Deacon.B.et al.1997, Global
et al.: Sporadic autism exomes reveal a highly interconnected protein network of de novo mutations. et al.: Patterns and rates of exonic de novo mutations in autism
et al., Determination of Dynamic Constitutive Equation with Temperature and Strain-rate Dependence for a Carbon Steel, Transactions of the Japan Society of Mechanical Engineers,
K T ¼ 0.9 is left unchanged from the de Pillis et al. [12] model, as we found no data supporting a different value. de Pillis et al. [12] took it originally from Ref. Table 4 of
For a brief history of the Fekete- Szeg¨o problem for class of starlike, convex, and close-to convex functions, see the recent paper by Srivastava et