[1]WEN Gui-hua,J IANG L i-jun,WEN Jun.Using locally estimated geodesic distances to improve Hessian local linear embedding[J].CAAI Transactions on Intelligent Systems,2008,3(5):429-435.
Copy

Using locally estimated geodesic distances to improve Hessian local linear embedding

CAAI Transactions on Intelligent Systems[ISSN 1673-4785/CN 23-1538/TP] Volume: 3 Number of periods: 2008 5 Page number: 429-435 Column: 学术论文—人工智能基础 Public date: 2008-10-25
Title:
Using locally estimated geodesic distances to improve Hessian local linear embedding
Author(s):
WEN Gui-hua1 ;  J IANG L i-jun2 ;  WEN Jun3
1. School of Computer Science and Engineering, South China University of Technology, Guangzhou 510641, China;
2. Department of ElectronicMaterial Science and Engineering, South China University of Technology, Gunagzhou 510641, China;
?3. School ofMathe2 matical Science, Hubei Insitute forNationalities, Enshi 445000, China
Keywords:
manifold learning;  Hessian transformation;  locally linear embedding;  geodesic distance
CLC:
TP391
DOI:
-
Abstract:
To dealwith highly curved data manifolds, the Hessian locally linear embedding (HLLE) algorithm was modified based on a locally estimated geodesic distance. It used the general concep tual framework of HLLE to guar2 antee correct setting of local isometry to an open connected subset. It emp loys the locally estimated geodesic dis2 tance instead of the Euclidean distance to determine the neighborhood of any point, so that it reduces the distorting influence of curvature of the data manifold on determining the neighborhood. This app roach can be regarded as the integration of a localmethod and a globalmethod, so that it has better performance and stability than HLLE, with only a slight increase in computational time. Experiments conducted on benchmark data sets verified etficioncy of the p roposed app roach.
References:
[ 1 ]TENENBAUM J B, De SILVA V, LANGFORD J C. A glob2 al geometric framework for nonlinear dimensionality reduction [ J ]. Science, 2000, 290 (12) : 231922323.
?[ 2 ]ROWEIS S T, SAUL L K. Nonlinear dimensionality reduc2 tion by locally linear embedding [ J ]. Science, 2000, 290 (12) : 231922323.
?[ 3 ]BELKIN M, N IYOGI P. Lap lacian eigenmap s for dimen2 sionality reduction and data rep resentation [ J ]. Neural Com2 puting, 2003, 15 (6) : 137321396.
[ 4 ]DONOHO D L, GR IMES C. Hessian eigenmap s: locally linear embedding, techniques for high2dimensional data[ J ]. PNAS, 2003, 100 (10) : 5912596.
[ 5 ]De R IDDER D, KOUROPTEVA O, OKUN O, et al. Super2 vised locally linear embedding [ C ] / /Artificial Neural Net2 works and Neural Information Processing. Istanbul, Turkey: Sp ringer, 2003: 3332341.
[ 6 ]KOUROPTEVA O, OKUN O, P IETIKA INEN M. Incremen2 tal locally linear embedding[ J ]. Pattern Recognition, 2005, 38 (10) : 1764217671
[ 7 ]De R IDDER D, LOOGM, REINDERSM J T. Local fisher embedding[C ] / /Proceedings of the 17 th International Con2 ference on Pattern Recognition. [ S. l. ] , 2004: 29522981
?[ 8 ]CHANG Hong, YENG Dityan. Robust locally linear embed2 ding[ J ]. Pattern Recognition, 2000, 39 (6) : 105321065.
[ 9 ]YANGL. Building k2connected neighborhood graphs for iso2 metric data embedding[ J ]. IEEE Transactions on Pattern A2 nalysis andMachine Intelligence, 2006, 28 (5) : 8272831.
[ 10 ]YANG L. Building connected neighborhood graphs for lo2 cally linear embedding[ C ] / /18 th International Conference on Pattern Recognition. Hong Kong, China, 2006: 16802 1683.
?[ 11 ]VAR IN I C, DEGENHARD A, NATTKEMPER T W. 　 ISOLLE: LLE with geodesic distance[ J ]. Neurocomputing, 2006, 69 (15) : 176821771.
?[ 12 ]王和勇,郑　杰,姚正安,李　磊. 基于聚类和改进距离的LLE方法在数据降维中的应用[ J ]. 计算机研究与发展, 2006, 43 (8) : 148521490.
?WANG Heyong, ZHENG J ie, YAO Zhengan, L ILei. Ap2 p lication of dimension reduction on using imp roved LLE based on clustering[ J ]. Journal of Computer Research and Development, 2006, 43 (8) : 148521490.
[ 13 ]WEN Guihua, J IANG L ijun, WEN Jun, N IGEL R S. Clustering2based nonlinear dimensionality reduction on manifold [ C ] / /Lecture Notes in Artificial Intelligence (LNA I). Sp ringer, 2006: 4442453.
?[ 14 ]文贵华、江丽君,文　军. 基于邻域优化的局部线性嵌入 [ J ]. 系统仿真学报, 2007, 19 (13) : 311923122.
?WEN Guihua, J IANGL ijun, WEN Jun. Locally linear em2 bedding based on op timization of neighborhood [ J ]. Journal of System Simulation, 2007, 19 (13) : 311923122.
?[ 15 ]WEN Guihua, J IANGL ijun , N IGEL R S. Using graph al2 gebra to op timize neighborhood for Isometric mapp ing [C ] / /20 th International Joint Conference on Artificial In2 telligence ( IJCA I207). India, 2007: 239822403.
[ 16 ] SAMKO O, MARSHALL A D, ROSIN P L. Selection of the op timal parameter value for the Isomap algorithm [ J ]. Pattern Recognition Letters, 2006, 27 (9) : 9682979.
[ 17 ]文贵华,江丽君,文　军. 邻域参数动态变化的局部线性嵌入[ J ]. 软件学报, 2008, 19 (7) : 166621673.
WEN Guihua, J IANGL ijun, WEN Jun. Dynamically deter2 mining neighborhood parameter for locally linear embedding [ J ]. Journal of Software, 2008, 19 (7) : 166621673.
?[ 18 ]WEN Guihua, J IANG L ijun. Globalizing local neighbor2 hood for locally linear embedding [ C ] / /Proceedings of 2006 IEEE International Conference on System, Man and Cybernetics. Taibei, China, 2006: 349123496.
?[ 19 ]De R IDDER D,LOOGM, REINDERSM J T. Local fisher embedding [ C ] / /Proceedings of the 17 th International Conference on Pattern Recognition. Cambridge, UK, 2004: 2952298.
[ 20 ] HONG Chang, YEUND Yang. Robust locally linear em2 bedding[ J ]. Pattern Recognition, 2003, 39 ( 6 ) : 10532 1065
Similar References:

Memo

-

Last Update: 2009-05-18

Copyright © CAAI Transactions on Intelligent Systems