?

Size: px
Start display at page:

Download "1 4 2 4 3 5 4? 7 5 9 6 10 7 11 8 13 9 16 10 17 11 19 12 20 13 21 2"

Transcription

1 78-2 (2002) p

2 ?

3 ( ) ( )? 3

4 1 N i p i log p i i p i log p i i N i q i N i p i log q i N i p i { ( log q i ) ( log p i ) } = N i p i log p i q i > 0 KL-divergence 0 < p i < 1, 0 < q i < 1, i p i = i q i = 1 i p i log p i autreasjkloiuxdecollkfdryuhhgreuiolpjhgfreasjoijnbvcxzswert N i p i log p i p i? p i n i i N = i n i p i = n i /N i? 1 () 4

5 i (1) (1) (1) (11) i n i = 1 i n i = 0 0 log 0 = 0 0 log log log 1 = 0 autreasjkloiuxdecollkfdryuhhgreuiolpjhgfreasjoijnbvcxzswert 1 1 (0?) N N 1 5

6 ( ) MDL (Minimum Description Length principle) MDL ( ) i p i log p i autreasjkloiuxdecollkfdryuhhgreuiolpjhgfreasjoijnbvcxzswert 1 autreasjkloiuxdecollkfdryuhhgreuiolpjhgfreasjoijnbvcxzswert O(1) O(N) N i p i q i q i i q i i p i q i y y i < y y i+1 i i q i ( ) q i exp ( (y i θ) 2 ) 2σ 2 i, i = y i+1 y i MDL θ P θ q i = P (i θ) 6

7 i n i log P (i θ) + (θ ) P, θ 4? θ θ θ = θ i π i θ i log π i 4? 399 MDL 10 2 n log n π(n) 1/n π(n) =? n n log n. log n log n + log log n + log log log n... log 5 π(n) n 4 θ π(θ) (prior distribution)

8 MDL π(n) π(n) π(n) = 1 n 2 10 π(n) π(n) 1/n Zipf 1/f n 6 2 π(n) MDL n L log L /L improper prior Lindley paradox 7 MDL 6 7 8

9 5 (b) (a) (b) (c) (d) (c) (b) (c) 8 ( )(c) overfit (d) 8? 9

10 6? ( ) AIC ( ) AIC i p i q i AIC

11 log P ( θ) - (θ ) P θ θ θ σ 2 y = f θ (x) P ({y i } θ) exp ( ) 1 2σ 2 (y i f θ (x i )) 2 1 2σ 2 (y i f θ (x i )) 2 + (θ ) i i 13 2 θ 33 4 i (y i f(x i )) ( ) 2. (cross validation) 3. (MDL,BIC, ) AIC 7? m m 2 m 13 σ 2 11

12 14 m m 2 m m 256 m = χ m m 250 AIC MDL χ 2? n = 1, 2,... n 2 1 m m 1 m m m 15 12

13 ( ) 16 ( ) Complexity is complicated. Complexity is a complicated matter.? 19 III 8 16 Knuth ( ) Knuth n, m 19 13

14 2 DNA DNA 3 TTG, CTT, CTC, Vol.4 No.6 pp (1989)

15 ( Jounal of Royal Statistical Society A ) MRI

16 H.S. ( ) ( ) 9 = + = + = + AIC MDL = / = / 16

17 = / AIC MDL 23 10? ( )

18 i p i log p i discommunication ( ) 18

19 11 ( ) 24? [ ] AIC MDL/BIC 26 19

20 (intention)

21 i p i log p i i p i i p i log p i representation re- 31 representation presentation representation representive presentation representation presentation representation recognition cognition 21

22 representation) (presentation) void solid void void solid CG (1997) 37 22

23 23

1 P2 P P3P4 P5P8 P9P10 P11 P12

1 P2 P P3P4 P5P8 P9P10 P11 P12 1 P2 P14 2 3 4 5 1 P3P4 P5P8 P9P10 P11 P12 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 & 11 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 1! 3 2 3! 4 4 3 5 6 I 7 8 P7 P7I P5 9 P5! 10 4!! 11 5 03-5220-8520

More information

…K…E…X„^…x…C…W…A…fi…l…b…g…‘†[…N‡Ì“‚¢−w‘K‡Ì‹ê™v’«‡É‡Â‡¢‡Ä

…K…E…X„^…x…C…W…A…fi…l…b…g…‘†[…N‡Ì“‚¢−w‘K‡Ì‹ê™v’«‡É‡Â‡¢‡Ä 2009 8 26 1 2 3 ARMA 4 BN 5 BN 6 (Ω, F, µ) Ω: F Ω σ 1 Ω, ϕ F 2 A, B F = A B, A B, A\B F F µ F 1 µ(ϕ) = 0 2 A F = µ(a) 0 3 A, B F, A B = ϕ = µ(a B) = µ(a) + µ(b) µ(ω) = 1 X : µ X : X x 1,, x n X (Ω) x 1,,

More information

チュートリアル:ノンパラメトリックベイズ

チュートリアル:ノンパラメトリックベイズ { x,x, L, xn} 2 p( θ, θ, θ, θ, θ, } { 2 3 4 5 θ6 p( p( { x,x, L, N} 2 x { θ, θ2, θ3, θ4, θ5, θ6} K n p( θ θ n N n θ x N + { x,x, L, N} 2 x { θ, θ2, θ3, θ4, θ5, θ6} log p( 6 n logθ F 6 log p( + λ θ F θ

More information

aisatu.pdf

aisatu.pdf 1 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71

More information

... 3... 3... 3... 3... 4... 7... 10... 10... 11... 12... 12... 13... 14... 15... 18... 19... 20... 22... 22... 23 2

... 3... 3... 3... 3... 4... 7... 10... 10... 11... 12... 12... 13... 14... 15... 18... 19... 20... 22... 22... 23 2 1 ... 3... 3... 3... 3... 4... 7... 10... 10... 11... 12... 12... 13... 14... 15... 18... 19... 20... 22... 22... 23 2 3 4 5 6 7 8 9 Excel2007 10 Excel2007 11 12 13 - 14 15 16 17 18 19 20 21 22 Excel2007

More information

PowerPoint プレゼンテーション

PowerPoint プレゼンテーション 0 1 2 3 4 5 6 1964 1978 7 0.0015+0.013 8 1 π 2 2 2 1 2 2 ( r 1 + r3 ) + π ( r2 + r3 ) 2 = +1,2100 9 10 11 1.9m 3 0.64m 3 12 13 14 15 16 17 () 0.095% 0.019% 1.29% (0.348%) 0.024% 0.0048% 0.32% (0.0864%)

More information

http://banso.cocolog-nifty.com/ 100 100 250 5 1 1 http://www.banso.com/ 2009 5 2 10 http://www.banso.com/ 2009 5 2 http://www.banso.com/ 2009 5 2 http://www.banso.com/ < /> < /> / http://www.banso.com/

More information

n=360 28.6% 34.4% 36.9% n=360 2.5% 17.8% 19.2% n=64 0.8% 0.3% n=69 1.7% 3.6% 0.6% 1.4% 1.9% < > n=218 1.4% 5.6% 3.1% 60.6% 0.6% 6.9% 10.8% 6.4% 10.3% 33.1% 1.4% 3.6% 1.1% 0.0% 3.1% n=360 0% 50%

More information

. R R D e R R 7 () r r R R () l t t R R 7 l () () R r rr r r n r n r r 3 6 r 88 R r 360 r = e t t = e r t rt rt, r t, r 3 t, r t R R R R R D = {e, r,

. R R D e R R 7 () r r R R () l t t R R 7 l () () R r rr r r n r n r r 3 6 r 88 R r 360 r = e t t = e r t rt rt, r t, r 3 t, r t R R R R R D = {e, r, 3 3 3 e X X X X X X . R R D e R R 7 () r r R R () l t t R R 7 l () () R r rr r r n r n r r 3 6 r 88 R r 360 r = e t t = e r t rt rt, r t, r 3 t, r t R R R R R D = {e, r, r, r 3, r, t, rt, r t, r 3 t, r

More information

FX ) 2

FX ) 2 (FX) 1 1 2009 12 12 13 2009 1 FX ) 2 1 (FX) 2 1 2 1 2 3 2010 8 FX 1998 1 FX FX 4 1 1 (FX) () () 1998 4 1 100 120 1 100 120 120 100 20 FX 100 100 100 1 100 100 100 1 100 1 100 100 1 100 101 101 100 100

More information

x T = (x 1,, x M ) x T x M K C 1,, C K 22 x w y 1: 2 2

x T = (x 1,, x M ) x T x M K C 1,, C K 22 x w y 1: 2 2 Takio Kurita Neurosceince Research Institute, National Institute of Advanced Indastrial Science and Technology takio-kurita@aistgojp (Support Vector Machine, SVM) 1 (Support Vector Machine, SVM) ( ) 2

More information

項 目

項 目 1 1 2 3 11 4 6 5 7,000 2 120 1.3 4,000 04 450 < > 5 3 6 7 8 9 4 10 11 5 12 45 6 13 E. 7 B. C. 14 15 16 17 18 19 20 21 22 23 8 24 25 9 27 2 26 6 27 3 1 3 3 28 29 30 9 31 32 33 500 1 4000 0 2~3 10 10 34

More information

01

01 2 0 0 7 0 3 2 2 i n d e x 0 7. 0 2. 0 3. 0 4. 0 8. 0 9. 1 0. 1 1. 0 5. 1 2. 1 3. 1 4. 1 5. 1 6. 1 7. 1 8. 1 9. 2 0. 2 1. 2 3. 2 4. 2 5. 2 6. O k h o t s k H a m a n a s u B e e f 0 2 http://clione-beef.n43.jp

More information

●70974_100_AC009160_KAPヘ<3099>ーシス自動車約款(11.10).indb

●70974_100_AC009160_KAPヘ<3099>ーシス自動車約款(11.10).indb " # $ % & ' ( ) * +, -. / 0 1 2 3 4 5 6 7 8 9 : ; < = >? @ A B C D E F G H I J K L M N O P Q R S T U V W X Y " # $ % & ' ( ) * + , -. / 0 1 2 3 4 5 6 7 8 9 : ; < = > ? @ A B

More information

(個別のテーマ) 薬剤に関連した医療事故

(個別のテーマ) 薬剤に関連した医療事故 - 67 - III - 68 - - 69 - III - 70 - - 71 - III - 72 - - 73 - III - 74 - - 75 - III - 76 - - 77 - III - 78 - - 79 - III - 80 - - 81 - III - 82 - - 83 - III - 84 - - 85 - - 86 - III - 87 - III - 88 - - 89

More information

(個別のテーマ) 放射線検査に関連した医療事故

(個別のテーマ) 放射線検査に関連した医療事故 - 131 - III - 132 - - 133 - III - 134 - - 135 - III - 136 - - 137 - III - 138 - - 139 - III - 140 - - 141 - III - 142 - - 143 - III - 144 - - 145 - III - 146 - - 147 - III - 148 - - 149 - III - 150 - -

More information

untitled

untitled 280 200 5 7,800 6 8,600 28 1 1 18 7 8 2 ( 31 ) 7 42 2 / / / / / / / / / / 1 3 (1) 4 5 3 1 1 1 A B C D 6 (1) -----) (2) -- ()) (3) ----(). ()() () ( )( )( )( ) ( ) ( )( )( )( ) () (). () ()() 7 () ( ) 1

More information

122011pp.139174 18501933

122011pp.139174 18501933 122011pp.139174 18501933 122011 1850 3 187912 3 1850 8 1933 84 4 1871 12 1879 5 2 1 9 15 1 1 5 3 3 3 6 19 9 9 6 28 7 7 4 1140 9 4 3 5750 58 4 3 1 57 2 122011 3 4 134,500,000 4,020,000 11,600,000 5 2 678.00m

More information

Microsoft Word - 映画『東京裁判』を観て.doc

Microsoft Word - 映画『東京裁判』を観て.doc 1 2 3 4 5 6 7 1 2008. 2 2010, 3 2010. p.1 4 2008 p.202 5 2008. p.228 6 2011. 7 / 2008. pp.3-4 1 8 1 9 10 11 8 2008, p.7 9 2011. p.41 10.51 11 2009. p. 2 12 13 14 12 2008. p.4 13 2008, p.7-8 14 2008. p.126

More information

日経テレコン料金表(2016年4月)

日経テレコン料金表(2016年4月) 1 2 3 4 8,000 15,000 22,000 29,000 5 6 7 8 36,000 42,000 48,000 54,000 9 10 20 30 60,000 66,000 126,000 166,000 50 100 246,000 396,000 1 25 8,000 7,000 620 2150 6,000 4,000 51100 101200 3,000 1,000 201

More information

73 p.1 22 16 2004p.152

73 p.1 22 16 2004p.152 1987 p.80 72 73 p.1 22 16 2004p.152 281895 1930 1931 12 28 1930 10 27 12 134 74 75 10 27 47.6 1910 1925 10 10 76 10 11 12 139 p.287 p.10 11 pp.3-4 1917 p.284 77 78 10 13 10 p.6 1936 79 15 15 30 80 pp.499-501

More information

29 2011 3 4 1 19 5 2 21 6 21 2 21 7 2 23 21 8 21 1 20 21 1 22 20 p.61 21 1 21 21 1 23

29 2011 3 4 1 19 5 2 21 6 21 2 21 7 2 23 21 8 21 1 20 21 1 22 20 p.61 21 1 21 21 1 23 29 2011 3 pp.55 86 19 1886 2 13 1 1 21 1888 1 13 2 3,500 3 5 5 50 4 1959 6 p.241 21 1 13 2 p.14 1988 p.2 21 1 15 29 2011 3 4 1 19 5 2 21 6 21 2 21 7 2 23 21 8 21 1 20 21 1 22 20 p.61 21 1 21 21 1 23 1

More information

() L () 20 1

() L () 20 1 () 25 1 10 1 0 0 0 1 2 3 4 5 6 2 3 4 9308510 4432193 L () 20 1 PP 200,000 P13P14 3 0123456 12345 1234561 2 4 5 6 25 1 10 7 1 8 10 / L 10 9 10 11 () ( ) TEL 23 12 7 38 13 14 15 16 17 18 L 19 20 1000123456

More information

308 ( ) p.121

308 ( ) p.121 307 1944 1 1920 1995 2 3 4 5 308 ( ) p.121 309 10 12 310 6 7 ( ) ( ) ( ) 50 311 p.120 p.142 ( ) ( ) p.117 p.124 p.118 312 8 p.125 313 p.121 p.122 p.126 p.128 p.156 p.119 p.122 314 p.153 9 315 p.142 p.153

More information

戦後の補欠選挙

戦後の補欠選挙 1 2 11 3 4, 1968, p.429., pp.140-141. 76 2005.12 20 14 5 2110 25 6 22 7 25 8 4919 9 22 10 11 12 13 58154 14 15 1447 79 2042 21 79 2243 25100 113 2211 71 113 113 29 p.85 2005.12 77 16 29 12 10 10 17 18

More information

2 2 3 4 5 5 2 7 3 4 6 1 3 4 7 4 2 2 2 4 2 3 3 4 5 1932 A p. 40. 1893 A p. 224, p. 226. 1893 B pp. 1 2. p. 3.

2 2 3 4 5 5 2 7 3 4 6 1 3 4 7 4 2 2 2 4 2 3 3 4 5 1932 A p. 40. 1893 A p. 224, p. 226. 1893 B pp. 1 2. p. 3. 1 73 72 1 1844 11 9 1844 12 18 5 1916 1 11 72 1 73 2 1862 3 1870 2 1862 6 1873 1 3 4 3 4 7 2 3 4 5 3 5 4 2007 p. 117. 2 2 3 4 5 5 2 7 3 4 6 1 3 4 7 4 2 2 2 4 2 3 3 4 5 1932 A p. 40. 1893 A p. 224, p. 226.

More information

「産業上利用することができる発明」の審査の運用指針(案)

「産業上利用することができる発明」の審査の運用指針(案) 1 1.... 2 1.1... 2 2.... 4 2.1... 4 3.... 6 4.... 6 1 1 29 1 29 1 1 1. 2 1 1.1 (1) (2) (3) 1 (4) 2 4 1 2 2 3 4 31 12 5 7 2.2 (5) ( a ) ( b ) 1 3 2 ( c ) (6) 2. 2.1 2.1 (1) 4 ( i ) ( ii ) ( iii ) ( iv)

More information

2009-9-2.indd

2009-9-2.indd Q No.1441 Q No.1442 Vol.33 NO.9 (2009) 30 (614) Q No.1443 Vol.33 NO.9 (2009) 31 (615) Q No.1444 Vol.33 NO.9 (2009) 32 (616) Vol.33 NO.9 (2009) 33 (617) Vol.33 NO.9 (2009) 34 (618) Q No.1445 Vol.33 NO.9

More information

みさき_1

みさき_1 2 3 4 5 6 7 1F 2F 8 9 10 11 17 18 19 20 21 22 23 24 31 25 26 27 28 29 30 8 1 2 3 4 5 6 7 6 8 7 16 7 8 9 10 11 12 14 15 13 17 23 Vol.41 8 6 20 11 7 15 7 23 7 7 7 16 23 23 8 13 18:00 22:00 722

More information

Vol..3 2010 2 10 2

Vol..3 2010 2 10 2 1 Vol..3 2010 2 10 2 Vol..3 2010 2 10 3 Vol..3 2010 2 10 4 Vol..3 2010 2 10 5 Vol..3 2010 2 10 6 Vol..3 2010 2 10 7 Vol..3 2010 2 10 8 Vol..3 2010 2 10 9 Vol..3 2010 2 10 10 Vol..3 2010 2 10 11 Vol..3

More information

H21_report

H21_report vol.4 1 2 6 10 14 18 20 22 24 25 1 2 2172 73 3 21925 926 21125 126 4 5 6 21629 630 7 21107 108 21127 128 8 9 10 21616 617 11 211026 1027 211213 1214 12 13 14 21713 714 15 2194 95 211031 111 16 17 18 19

More information

untitled

untitled Vol.27 1 Vol.27 2 Vol.27 3 Vol.27 4 Vol.27 5 Vol.27 6 Vol.27 7 Vol.27 8 Vol.27 9 Vol.27 10 11 Vol.27 Vol.27 12 Vol.27 13 Vol.27 14 Vol.27 15 Vol.27 16 Vol.27 17 Vol.27 2007 10 29 18 http://www.nira.or.jp/index.html

More information