: W, k : C 1,, C k 1. W D ii = j W ij D 2. W, D L = I D 1/2 W D 1/2 L 3. L, k U 4. U k-means C 3: 2: 3. ( ) k-means HITS k-means k-mean

Size: px

Start display at page:

Download ": W, k : C 1,, C k 1. W D ii = j W ij D 2. W, D L = I D 1/2 W D 1/2 L 3. L, k U 4. U k-means C 3: 2: 3. ( ) k-means 10 1 0 688 3.1 HITS k-means k-mean"

いおりさくもと
9 years ago
Views:

1 人工知能学会研究会資料 SIG-FIN Attempt Diversification by Clustering of Investment Trusts 1 Takumasa Sakakibara 2 Tohgoroh Matsui 1 Atsuko Mutoh 1 Nobuhiro Inuduka 1 Department of Computer Science and Engineering, Graduate School of Engineering, Nagoya Institute of Technology 2 Chubu University Investing in single or similar financial instruments is dangerous from the viewpoint of risk. In order to realize wide range of investment, make the clustering focuses on stocks companies that Investment Trusts are investing in this study. And I examine a wide range of investment can help to diversification [1, 2] : [email protected] V f v i, v j V f : G = (V, E) : V = V f : w(v i, v j ) = v i, v j : E = {(v i, v j) w(v i, v j) > 0} V f V f 1: Yahoo! MORNINGSTAR 10 Python HTML

: W, k : C 1,, C k 1. W D ii = j W ij D 2. W, D L = I D 1/2 W D 1/2 L 3. L, k U 4. U k-means C 3: 2: 3. ( ) k-means 10 1 0 688 3.

2 : W, k : C 1,, C k 1. W D ii = j W ij D 2. W, D L = I D 1/2 W D 1/2 L 3. L, k U 4. U k-means C 3: 2: 3. ( ) k-means HITS k-means k-means k i=1 x C i (dist(x, c i)) 2 k x C c dist(x, y) x y ( ) 1 N 1 2 [6] [3, 4, 5] ( ) k-means 1 W W ij = w(v i, v j ) v w(v i, v j ) v i v j 3 HITS HITS [7] authority Web hub Web Web Web Web Web Web 3.2 n, m = m n(n 1)/2 2

3 1: ( ) 5 k-means (188) 10 k-means (118) 30 k-means (65) (217) (117) (101) (52) (9) (11) ( 496) k-means 4 k-means ( ) () 4 1 HITS k-means HITS k-means 5 5 k-means 3.3 k k 4: k-means : 5 5: : 5 6 HITS k-means (5,10,20,30,50) HITS k-means 5 1 k/5 k-means 4k/5 4k/5 k 4. 3

4 6: 10 k-means 6 ( ) k-means k-means kmeans ( 5 7: ( ) 8: ( ) 9: ( ) 10: ( ) 4

5 11: 2012 ( ) : ( ) 13: ( ) ) k-means [1],,,, 75, 5M-9 (2013) [2],,,, 10 (SIG-FIN), SIG-FIN , pp (2013) [3] U. von Luxburg, A tutorial on spectral clustering, Stat. Comp. Vol. 17, Issue 4, (2007) [4] Alexandros Karatzoglou, Alex Smola, Kurt Hornik, Achim Zeileis, kernlab -An S4 Package for Kernel Methods in R [5] Jianbo Shi and Jitendra Malik, Normalized Cuts and Image Segmentation, IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLI- GENCE, VOL. 22, NO. 8, AUGUST 2000 [6], (1),, pp.59-65, (2003) [7] Jon Kleinberg, Authoritative sources in a hyperlinked environment, ACM-SIAM Symposium on Discrete Algorithms,pp (1998) ( 323) 1( 272) 120 MHAM UFJ F( ) : JATOPIX DIAM 225 5

6 UFJ : :MIP[MonthlyInvestmentPlan] BR : : : SRI: (3 ) GS TOPIX UFJ : 21 3 TOPIX :DCTOPIX : 2101 : 225 MHAMTOPIX emaxis 225 FVol.4( ) emaxistopix GS ( ) GS ( ) : ( ) ( ) MHAM J:K2000 J :DCJ : ( ) :NO.1 : IBJITM SMTTOPIX DIAM : II BR F 225 :Funds-i 225 ING : M A MHAM : GS 225 : SRI : F : BR ( SF) 225 : : : : E : 30 : : PB : : :RAO( ) GS: 2( 34)) JPM JF : : : - - :jnext : 30 DIAM SRI : SBI :jrevive : JPM 96 JASDAQ JPM ( ) JASDAQ-TOP20 1 JF JF P 3( 10) JPM JF F JPM (3 ) 4( 5) : SBI :jcool : 5( 2) ( 10)D 6

1 Fig. 1 Extraction of motion,.,,, 4,,, 3., 1, 2. 2.,. CHLAC,. 2.1,. (256 ).,., CHLAC. CHLAC, HLAC. 2.3 (HLAC ) r,.,. HLAC. N. 2 HLAC Fig. 2

1 Fig. 1 Extraction of motion,.,,, 4,,, 3., 1, 2. 2.,. CHLAC,. 2.1,. (256 ).,., CHLAC. CHLAC, HLAC. 2.3 (HLAC ) r,.,. HLAC. N. 2 HLAC Fig. 2 CHLAC 1 2 3 3,. (CHLAC), 1).,.,, CHLAC,.,. Suspicious Behavior Detection based on CHLAC Method Hideaki Imanishi, 1 Toyohiro Hayashi, 2 Shuichi Enokida 3 and Toshiaki Ejima 3 We have proposed a method for