Presentation 2020-01-09
Statistical Learning Theory of Data changed in Value
Satoshi Kataoka,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) In Statistical Learning Theory, the accuracy of inference is evaluated by generalization loss or free energy of true distribution. When data is changed in value because of data preprocessing, record miss and so on, the distribution used by inference changes too and we did not know behavor of generalization loss of the distribution before changing. Newly in this paper, we derive theoretical behavor of generalization loss of the distribution before changing and check the correctness by numerical experiment.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) generalization loss / statistical learning theorey / data changed in value
Paper # IBISML2019-21
Date of Issue 2020-01-02 (IBISML)

Conference Information
Committee IBISML
Conference Date 2020/1/9(2days)
Place (in Japanese) (See Japanese page)
Place (in English) ISM
Topics (in Japanese) (See Japanese page)
Topics (in English) Machine learning, etc.
Chair Hisashi Kashima(Kyoto Univ.)
Vice Chair Masashi Sugiyama(Univ. of Tokyo) / Koji Tsuda(Univ. of Tokyo)
Secretary Masashi Sugiyama(Nagoya Inst. of Tech.) / Koji Tsuda(AIST)
Assistant Tomoharu Iwata(NTT) / Shigeyuki Oba(Kyoto Univ.)

Paper Information
Registration To Technical Committee on Infomation-Based Induction Sciences and Machine Learning
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Statistical Learning Theory of Data changed in Value
Sub Title (in English)
Keyword(1) generalization loss
Keyword(2) statistical learning theorey
Keyword(3) data changed in value
1st Author's Name Satoshi Kataoka
1st Author's Affiliation Tokyo Institute of Technology(Titech)
Date 2020-01-09
Paper # IBISML2019-21
Volume (vol) vol.119
Number (no) IBISML-360
Page pp.pp.25-30(IBISML),
#Pages 6
Date of Issue 2020-01-02 (IBISML)