Presentation 2011-03-28
Regression:Possible : One dimensional elimination by hyperplane fitting
Jun FUJIKI, Shotaro AKAHO,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) In this paper, N-1 dimensional hyperplane fitting for N dimensional data is investigated. Firstly, many kinds of regression and/or principal component analysis are systematically classified. Secondly, it is proved that useful property that the global optimum pass through N data points for affine hyperplane regression, and N-1 data points for linear hyperplane regression for some kinds of regression. To evaluate the robustness of the estimated hyperplane by colinearlity of the data, the contribution rate defined in original principal component analysis is extended to many kinds of regressions based on L_2-norm.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) regression / hyperplane fitting / L_p-norm / global optimum / conminatorial optimization
Paper # IBISML2010-116
Date of Issue

Conference Information
Committee IBISML
Conference Date 2011/3/21(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To Information-Based Induction Sciences and Machine Learning (IBISML)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Regression:Possible : One dimensional elimination by hyperplane fitting
Sub Title (in English)
Keyword(1) regression
Keyword(2) hyperplane fitting
Keyword(3) L_p-norm
Keyword(4) global optimum
Keyword(5) conminatorial optimization
1st Author's Name Jun FUJIKI
1st Author's Affiliation The National Institute of Advanced Industrial Science and Technology (AIST)()
2nd Author's Name Shotaro AKAHO
2nd Author's Affiliation The National Institute of Advanced Industrial Science and Technology (AIST)
Date 2011-03-28
Paper # IBISML2010-116
Volume (vol) vol.110
Number (no) 476
Page pp.pp.-
#Pages 8
Date of Issue