Learning nonlinear multiregression networks based on evolutionary computation

Kwong Sak LEUNG, Man Leung WONG, Wai LAM, Zhenyuan WANG, Kebin XU

Research output: Journal PublicationsJournal Article (refereed)Researchpeer-review

22 Citations (Scopus)

Abstract

This paper describes a novel knowledge discovery and data mining framework dealing with nonlinear interactions among domain attributes. Our network-based model provides an effective and efficient reasoning procedure to perform prediction and decision making. Unlike many existing paradigms based on linear models, the attribute relationship in our framework is represented by nonlinear nonnegative multiregressions based on the Choquet integral. This kind of multiregression is able to model a rich set of nonlinear interactions directly. Our framework involves two layers. The outer layer is a network structure consisting of network elements as its components, while the inner layer is concerned with a particular network element modeled by Choquet integrals. We develop a fast double optimization algorithm (FDOA) for learning the multiregression coefficients of a single network element. Using this local learning component and multiregression-residual-cost evolutionary programming (MRCEP), we propose a global learning algorithm, called MRCEP-FDOA, for discovering the network structures and their elements from databases. We have conducted a series of experiments to assess the effectiveness of our algorithm and investigate the performance under different parameter combinations, as well as sizes of the training data sets. The empirical results demonstrate that our framework can successfully discover the target network structure and the regression coefficients.
Original languageEnglish
Pages (from-to)630-644
Number of pages15
JournalIEEE Transactions on Systems, Man, and Cybernetics, Part B: Cybernetics
Volume32
Issue number5
DOIs
Publication statusPublished - 1 Jan 2002

Fingerprint

Nonlinear networks
Evolutionary algorithms
Data mining
Learning algorithms
Costs
Decision making
Experiments

Keywords

  • Choquet integrals
  • Data mining
  • Evolutionary computation
  • Nonlinear multiregression networks

Cite this

@article{3428de6c0a24492f9eb2bcb38c480f80,
title = "Learning nonlinear multiregression networks based on evolutionary computation",
abstract = "This paper describes a novel knowledge discovery and data mining framework dealing with nonlinear interactions among domain attributes. Our network-based model provides an effective and efficient reasoning procedure to perform prediction and decision making. Unlike many existing paradigms based on linear models, the attribute relationship in our framework is represented by nonlinear nonnegative multiregressions based on the Choquet integral. This kind of multiregression is able to model a rich set of nonlinear interactions directly. Our framework involves two layers. The outer layer is a network structure consisting of network elements as its components, while the inner layer is concerned with a particular network element modeled by Choquet integrals. We develop a fast double optimization algorithm (FDOA) for learning the multiregression coefficients of a single network element. Using this local learning component and multiregression-residual-cost evolutionary programming (MRCEP), we propose a global learning algorithm, called MRCEP-FDOA, for discovering the network structures and their elements from databases. We have conducted a series of experiments to assess the effectiveness of our algorithm and investigate the performance under different parameter combinations, as well as sizes of the training data sets. The empirical results demonstrate that our framework can successfully discover the target network structure and the regression coefficients.",
keywords = "Choquet integrals, Data mining, Evolutionary computation, Nonlinear multiregression networks",
author = "LEUNG, {Kwong Sak} and WONG, {Man Leung} and Wai LAM and Zhenyuan WANG and Kebin XU",
year = "2002",
month = "1",
day = "1",
doi = "10.1109/TSMCB.2002.1033182",
language = "English",
volume = "32",
pages = "630--644",
journal = "IEEE Transactions on Systems, Man, and Cybernetics, Part B: Cybernetics",
issn = "1083-4419",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "5",

}

Learning nonlinear multiregression networks based on evolutionary computation. / LEUNG, Kwong Sak; WONG, Man Leung; LAM, Wai; WANG, Zhenyuan; XU, Kebin.

In: IEEE Transactions on Systems, Man, and Cybernetics, Part B: Cybernetics, Vol. 32, No. 5, 01.01.2002, p. 630-644.

Research output: Journal PublicationsJournal Article (refereed)Researchpeer-review

TY - JOUR

T1 - Learning nonlinear multiregression networks based on evolutionary computation

AU - LEUNG, Kwong Sak

AU - WONG, Man Leung

AU - LAM, Wai

AU - WANG, Zhenyuan

AU - XU, Kebin

PY - 2002/1/1

Y1 - 2002/1/1

N2 - This paper describes a novel knowledge discovery and data mining framework dealing with nonlinear interactions among domain attributes. Our network-based model provides an effective and efficient reasoning procedure to perform prediction and decision making. Unlike many existing paradigms based on linear models, the attribute relationship in our framework is represented by nonlinear nonnegative multiregressions based on the Choquet integral. This kind of multiregression is able to model a rich set of nonlinear interactions directly. Our framework involves two layers. The outer layer is a network structure consisting of network elements as its components, while the inner layer is concerned with a particular network element modeled by Choquet integrals. We develop a fast double optimization algorithm (FDOA) for learning the multiregression coefficients of a single network element. Using this local learning component and multiregression-residual-cost evolutionary programming (MRCEP), we propose a global learning algorithm, called MRCEP-FDOA, for discovering the network structures and their elements from databases. We have conducted a series of experiments to assess the effectiveness of our algorithm and investigate the performance under different parameter combinations, as well as sizes of the training data sets. The empirical results demonstrate that our framework can successfully discover the target network structure and the regression coefficients.

AB - This paper describes a novel knowledge discovery and data mining framework dealing with nonlinear interactions among domain attributes. Our network-based model provides an effective and efficient reasoning procedure to perform prediction and decision making. Unlike many existing paradigms based on linear models, the attribute relationship in our framework is represented by nonlinear nonnegative multiregressions based on the Choquet integral. This kind of multiregression is able to model a rich set of nonlinear interactions directly. Our framework involves two layers. The outer layer is a network structure consisting of network elements as its components, while the inner layer is concerned with a particular network element modeled by Choquet integrals. We develop a fast double optimization algorithm (FDOA) for learning the multiregression coefficients of a single network element. Using this local learning component and multiregression-residual-cost evolutionary programming (MRCEP), we propose a global learning algorithm, called MRCEP-FDOA, for discovering the network structures and their elements from databases. We have conducted a series of experiments to assess the effectiveness of our algorithm and investigate the performance under different parameter combinations, as well as sizes of the training data sets. The empirical results demonstrate that our framework can successfully discover the target network structure and the regression coefficients.

KW - Choquet integrals

KW - Data mining

KW - Evolutionary computation

KW - Nonlinear multiregression networks

UR - http://commons.ln.edu.hk/sw_master/2315

U2 - 10.1109/TSMCB.2002.1033182

DO - 10.1109/TSMCB.2002.1033182

M3 - Journal Article (refereed)

VL - 32

SP - 630

EP - 644

JO - IEEE Transactions on Systems, Man, and Cybernetics, Part B: Cybernetics

JF - IEEE Transactions on Systems, Man, and Cybernetics, Part B: Cybernetics

SN - 1083-4419

IS - 5

ER -