Invexity of the minimum error entropy criterion

Mujahid Syed; Panos Pardalos; Jose Principe

doi:10.1109/LSP.2013.2283425

Invexity of the minimum error entropy criterion

Mujahid Syed, Panos Pardalos, Jose Principe

Research output: Contribution to journal › Article › peer-review

14 Scopus citations

Abstract

In this letter, optimization properties of Minimization of Error Entropy (MEE) and Minimization of Error Entropy with Fiducial points (MEEF) are presented. It is proved that by varying the kernel parameter of the MEE and/or MEEF objective function, the resulting problem, in general leads to an invex problem. Furthermore, for certain values of the kernel parameter it is shown that the problems may transform to convex or pseudo-convex problems.

Original language	English
Article number	6619422
Pages (from-to)	1159-1162
Number of pages	4
Journal	IEEE Signal Processing Letters
Volume	20
Issue number	12
DOIs	https://doi.org/10.1109/LSP.2013.2283425
State	Published - 2013
Externally published	Yes

Keywords

Entropic learning
Error entropy
Information potential
Information theoretic learning

ASJC Scopus subject areas

Signal Processing
Electrical and Electronic Engineering
Applied Mathematics

Access to Document

10.1109/LSP.2013.2283425

Cite this

@article{d52f700513d7493791cce4d643ade30b,

title = "Invexity of the minimum error entropy criterion",

abstract = "In this letter, optimization properties of Minimization of Error Entropy (MEE) and Minimization of Error Entropy with Fiducial points (MEEF) are presented. It is proved that by varying the kernel parameter of the MEE and/or MEEF objective function, the resulting problem, in general leads to an invex problem. Furthermore, for certain values of the kernel parameter it is shown that the problems may transform to convex or pseudo-convex problems.",

keywords = "Entropic learning, Error entropy, Information potential, Information theoretic learning",

author = "Mujahid Syed and Panos Pardalos and Jose Principe",

year = "2013",

doi = "10.1109/LSP.2013.2283425",

language = "English",

volume = "20",

pages = "1159--1162",

journal = "IEEE Signal Processing Letters",

issn = "1070-9908",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

number = "12",

}

TY - JOUR

T1 - Invexity of the minimum error entropy criterion

AU - Syed, Mujahid

AU - Pardalos, Panos

AU - Principe, Jose

PY - 2013

Y1 - 2013

N2 - In this letter, optimization properties of Minimization of Error Entropy (MEE) and Minimization of Error Entropy with Fiducial points (MEEF) are presented. It is proved that by varying the kernel parameter of the MEE and/or MEEF objective function, the resulting problem, in general leads to an invex problem. Furthermore, for certain values of the kernel parameter it is shown that the problems may transform to convex or pseudo-convex problems.

AB - In this letter, optimization properties of Minimization of Error Entropy (MEE) and Minimization of Error Entropy with Fiducial points (MEEF) are presented. It is proved that by varying the kernel parameter of the MEE and/or MEEF objective function, the resulting problem, in general leads to an invex problem. Furthermore, for certain values of the kernel parameter it is shown that the problems may transform to convex or pseudo-convex problems.

KW - Entropic learning

KW - Error entropy

KW - Information potential

KW - Information theoretic learning

UR - https://www.scopus.com/pages/publications/84885675375

U2 - 10.1109/LSP.2013.2283425

DO - 10.1109/LSP.2013.2283425

M3 - Article

AN - SCOPUS:84885675375

SN - 1070-9908

VL - 20

SP - 1159

EP - 1162

JO - IEEE Signal Processing Letters

JF - IEEE Signal Processing Letters

IS - 12

M1 - 6619422

ER -

Invexity of the minimum error entropy criterion

Abstract

Keywords

ASJC Scopus subject areas

Access to Document

Fingerprint

Cite this