[5] Yahoo! Yahoo! (SVM) 3 F 7 7 (SVM) 3 F 6 0
1 1 2 3 2.1.................. 3 2.2.......... 6 3 7 3.1......................... 7 3.1.1 ALAGIN................ 7 3.1.2 (SVM)........................... 8 3.1.3 10................... 10 3.1.4.......................... 10 3.1.5 F....................... 11 3.2............................ 12 3.3............................ 13 4 16 4.1............................ 16 4.2............................ 17 5 20 6 21 1
3.1......................... 15 4.1.......................... 17 4.2........................... 18 4.3 F....................... 18 4.4 F.............. 19 4.5 F......... 19 4.6 F......... 19 2
2.1....................... 4 3.1 -................... 7 3.2 -................ 8 3.3 ALAGIN.......... 8 3.4............................... 9 3.5 10....................... 11 3.6.............................. 12 3.7...................... 13 3.8 2................................... 15 4.1...................... 16 4.2...................... 17 3
1 [5] Yahoo! Yahoo! Yahoo 1
Yahoo! 2
2 2. 2.1 [1] ALAGIN [10] 2.1 ALAGIN F 0.86 F 0.91 F 6 7. [3] 3
2.1: 8 5 4 F 1. n 2. 3. 4. 4
1 0.06 0.4 [2] [6] Twitter 0.19 0.94 4 1. 2. 3. 4. 4 5
2.2 [5] [4] Yahoo! Yahoo! 5 5 ( URL URL) URL URL ( ) 6
3 3.1 3.1.1 ALAGIN ALAGIN [10] 6 URL [11] B A A B 3.1 3.2 A B -, EB -, -, -, -, -, - PTSD, -, -, 3.1: - ( ) ( ) 7
A B -, -, -, -, -, -, -, -, 2 - ND, - 3.2: - A B B A A B A B B A 3.3 6 100 58,700,000 3.3: ALAGIN 3.1.2 (SVM) SVM 2 [7] 2 ( ) ( 3.4 ) ( ) 3.1 2 [8],[9] 8
3.4: ( l ) f(x) = sgn α i y i K(x i, x) + b i=1 b = max i,y i = 1b i + min i,yi =1b i 2 l b i = α j y j K(x j, x i ) j=1 (3.1) x ( ) x i y i (i = 1,..., l, y i {1, 1}) sgn sgn(x) = 1 (x 0) (3.2) 1 (otherwise) α i (3.4) (3.5) (3.3) L(α) L(α) = l α i 1 l α i α j y i y j K(x i, x j ) (3.3) 2 i=1 i,j=1 0 α i C (i = 1,..., l) (3.4) 9
l α i y i = 0 (3.5) i=1 K (3.6) K(x, y) = (x y + 1) d (3.6) C, d C,d 1 α i > 0 x i (3.1) 2 3 [8] N (N(N-1)/2 ) 2 N(N-1)/2 2 2 3.1.3 10 10 10 10 10 10 3.5 10 3.1.4 10
3.5: 10 3.6 3.1.5 F (P) (R) F (F) T T T T P = T T T (3.7) 11
3.6: R = T T T (3.8) F = ( 1 P + 1 R 2 ) 1 (3.9) F 3.2 ALAGIN ALAGIN URL URL B A 3.7 12
3.7: ChaSen[12] SVM 10 3.3 7 13
( ) 14
ChaSen SVM 7 2 2 3.8 3.8: 2 3.1 3.1: 676 IT 1,253 15
4 4.1 B A ALAGIN 746,432 1,000 4.1 4.1: 1,000 10 4.1 16
4.1: F 0.75 0.39 0.51 0.70 0.74 0.72 0.03 1.00 0.06 F 7 4.2 9,908 300 300 4.2 4.2: 300 4.2 4.3 4.4 4.5 4.6 F 6 F 5 17
4.2: 94 91 73 56 22 35 39 4.3: F 0.75 0.83 0.48 0.52 0.50 0.39 0.34 0.53 0.14 0.69 0.75 0.23 0.45 0.47 0.47 0.34 0.44 0.31 0.45 0.89 0.21 0.51 0.63 0.32 F 8 18
4.4: F F 0.82 (65/79) 0.69 (65/94) 0.75 0.60 (33/55) 0.45 (33/73) 0.52 0.37 (7/19) 0.32 (7/22) 0.34 0.72 (26/36) 0.67 (26/39) 0.69 0.50 (37/74) 0.41 (37/91) 0.45 0.62 (13/21) 0.23 (13/56) 0.34 0.67 (12/18) 0.34 (12/35) 0.45 0.61 0.44 0.51 4.5: F F 0.73 (91/125) 0.97 (91/94) 0.83 0.48 (39/81) 0.53 (39/73) 0.51 0.42 (16/38) 0.73 (16/22) 0.53 0.63 (36/57) 0.92 (36/39) 0.75 0.30 (91/300) 1.00 (91/91) 0.47 0.69 (18/26) 0.32 (18/56) 0.44 0.8 (35/44) 1.00 (35/35) 0.89 0.58 0.78 0.63 4.6: F F 0.31 (94/300) 1.00 (94/94) 0.48 0.24 (73/300) 1.00 (73/73) 0.39 0.07 (22/300) 1.00 (22/22) 0.14 0.13 (39/300) 1.00 (39/39) 0.23 0.30 (91/300) 1.00 (91/91) 0.47 0.19 (56/300) 1.00 (56/56) 0.31 0.12 (35/300) 1.00 (35/35) 0.21 0.20 1.00 0.32 19
5 8 [4] 7 [6] 20
6 F 5 7 F 5 6 8 21
C C 22
[1].. 19, pp.592-595, 2013. [2] Twitter., (NLC), NLC2014-1, pp.1-6, 2014. [3],,.. 18, pp.303-306, 2012. [4],.. 50 ( 72 ), pp. 4-853 - 4-854, 2010. [5].., 2014. [6],,..., pp.53-60, 2004. [7].. 2001-NL-144, pp.113-120, 2001. [8] Taku Kudoh TinySVM: Support Vector Machines, http:// www.chasen.org/taku/software/tinysvm/, 2000. [9] Nello Cristianini, John Shawe-Taylor. An Introduction to Support Vector Machines and Other Kernel-based Learning Methods, Cambridge University Press, 2000. [10] ALAGIN : https://alaginrc.nict.go.jp/. 23
[11] :, https://alaginrc.nict.go.jp/, ( ) MASTAR. [12] ChaSen: http://chasen-legacy.sourceforge.jp/. 24