Information geometric perspective of modal linear regression

Keishi Sando, Shotaro Akaho, Noboru Murata, Hideitsu Hino

Research output: Chapter in Book/Report/Conference proceedingConference contribution

1 Citation (Scopus)

Abstract

Modal linear regression (MLR) is a standard method for modeling the conditional mode of a response variable using a linear combination of explanatory variables. It is effective when dealing with response variables with an asymmetric, multi-modal distribution. Because of the nonparametric nature of MLR, it is difficult to construct a statistical model manifold in the sense of information geometry. In this work, a model manifold is constructed using observations instead of explicit parametric models. We also propose a method for constructing a data manifold based on an empirical distribution. The em algorithm, which is a geometric formulation of the EM algorithm, of MLR is shown to be equivalent to the conventional EM algorithm of MLR.

Original languageEnglish
Title of host publicationNeural Information Processing - 25th International Conference, ICONIP 2018, Proceedings
EditorsLong Cheng, Seiichi Ozawa, Andrew Chi Sing Leung
PublisherSpringer Verlag
Pages535-545
Number of pages11
ISBN (Print)9783030041816
DOIs
Publication statusPublished - 2018
Event25th International Conference on Neural Information Processing, ICONIP 2018 - Siem Reap, Cambodia
Duration: 2018 Dec 132018 Dec 16

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume11303 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

Other25th International Conference on Neural Information Processing, ICONIP 2018
Country/TerritoryCambodia
CitySiem Reap
Period18/12/1318/12/16

Keywords

  • EM algorithm
  • Information geometry
  • Modal linear regression

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Fingerprint

Dive into the research topics of 'Information geometric perspective of modal linear regression'. Together they form a unique fingerprint.

Cite this