Presentation 1994/1/28
Feature Extraction from High-Dimensional Data-Evaluation of the Effect of smoothing-
Senya Kiyasu, Sadao Fujimura,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) In order to efficiently and accurately obtain the results from high-dimensional data,significant features should be extracted before processing.We already proposed a feature extraction method for significance weighted supervised classification.But the weighting factors for each dimensions,which are the results of the feature extraction,were not smooth enough along the dimension.In order to reduce the influence of random noise,we smoothed the data by moving average method and used for feature extraction.The extracted weighting factors were much smoother by the effect of noise reduction.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) feature extraction / high-dimensional data / smoothing / noise / classification
Paper # SANE93-79
Date of Issue

Conference Information
Committee SANE
Conference Date 1994/1/28(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 Space, Aeronautical and Navigational Electronics (SANE)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Feature Extraction from High-Dimensional Data-Evaluation of the Effect of smoothing-
Sub Title (in English)
Keyword(1) feature extraction
Keyword(2) high-dimensional data
Keyword(3) smoothing
Keyword(4) noise
Keyword(5) classification
1st Author's Name Senya Kiyasu
1st Author's Affiliation Department of Mathematical Engineering and Information Physics, Faculty of Engineering,University of Tokyo()
2nd Author's Name Sadao Fujimura
2nd Author's Affiliation Department of Mathematical Engineering and Information Physics, Faculty of Engineering,University of Tokyo
Date 1994/1/28
Paper # SANE93-79
Volume (vol) vol.93
Number (no) 453
Page pp.pp.-
#Pages 6
Date of Issue