A Micro-Extended Belief Rule-Based System for Big Data Multi-Class Classification Problems

Longhao Yang, J. Liu, Yingming Wang, Luis Martinez

Research output: Contribution to journalArticle

Abstract

Big data classification problems have drawn great attention from diverse fields and many classifiers have been developed. Among those classifiers, the extended belief rule-based system (EBRBS) has shown its potential in both big data and multi-class situations while time complexity and computing efficiency are two challenging issues to be handled in EBRBS. As such, three improvements of EBRBS are proposed firstly in the present paper to decrease the time complexity and computing efficiency of EBRBS for multi-class classification under the assumption of large amount of data, including the strategy to skip rule weight calculation, a simplified evidential reasoning algorithm, and the domain division-based rule reduction method. This turns out to be a micro version of the EBRBS, classed Micro-EBRBS. Moreover, one of commonly used cluster computing, named Apache Spark, is then applied to implement the parallel rule generation and inference schemes of the Micro-EBRBS for big data multi- class classification problems. The comparative analyses of experimental studies demonstrate that the Micro-EBRBS not only can obtain a desired accuracy, but also has the comparatively better time complexity and computing efficiency than some popular classifiers, especially for multi-class classification problems.
LanguageEnglish
Pages1-21
JournalIEEE Transactions on Systems, Man, and Cybernetics: Systems
DOIs
Publication statusAccepted/In press - 28 Oct 2018

Fingerprint

Knowledge based systems
Classifiers
Cluster computing
Big data
Electric sparks

Cite this

@article{13509e5064774921bfee4d56816ef7a4,
title = "A Micro-Extended Belief Rule-Based System for Big Data Multi-Class Classification Problems",
abstract = "Big data classification problems have drawn great attention from diverse fields and many classifiers have been developed. Among those classifiers, the extended belief rule-based system (EBRBS) has shown its potential in both big data and multi-class situations while time complexity and computing efficiency are two challenging issues to be handled in EBRBS. As such, three improvements of EBRBS are proposed firstly in the present paper to decrease the time complexity and computing efficiency of EBRBS for multi-class classification under the assumption of large amount of data, including the strategy to skip rule weight calculation, a simplified evidential reasoning algorithm, and the domain division-based rule reduction method. This turns out to be a micro version of the EBRBS, classed Micro-EBRBS. Moreover, one of commonly used cluster computing, named Apache Spark, is then applied to implement the parallel rule generation and inference schemes of the Micro-EBRBS for big data multi- class classification problems. The comparative analyses of experimental studies demonstrate that the Micro-EBRBS not only can obtain a desired accuracy, but also has the comparatively better time complexity and computing efficiency than some popular classifiers, especially for multi-class classification problems.",
author = "Longhao Yang and J. Liu and Yingming Wang and Luis Martinez",
year = "2018",
month = "10",
day = "28",
doi = "10.1109/TSMC.2018.2872843",
language = "English",
pages = "1--21",
journal = "IEEE Transactions on Systems, Man, and Cybernetics: Systems",
issn = "2168-2216",

}

A Micro-Extended Belief Rule-Based System for Big Data Multi-Class Classification Problems. / Yang, Longhao; Liu, J.; Wang, Yingming; Martinez, Luis.

In: IEEE Transactions on Systems, Man, and Cybernetics: Systems, 28.10.2018, p. 1-21.

Research output: Contribution to journalArticle

TY - JOUR

T1 - A Micro-Extended Belief Rule-Based System for Big Data Multi-Class Classification Problems

AU - Yang, Longhao

AU - Liu, J.

AU - Wang, Yingming

AU - Martinez, Luis

PY - 2018/10/28

Y1 - 2018/10/28

N2 - Big data classification problems have drawn great attention from diverse fields and many classifiers have been developed. Among those classifiers, the extended belief rule-based system (EBRBS) has shown its potential in both big data and multi-class situations while time complexity and computing efficiency are two challenging issues to be handled in EBRBS. As such, three improvements of EBRBS are proposed firstly in the present paper to decrease the time complexity and computing efficiency of EBRBS for multi-class classification under the assumption of large amount of data, including the strategy to skip rule weight calculation, a simplified evidential reasoning algorithm, and the domain division-based rule reduction method. This turns out to be a micro version of the EBRBS, classed Micro-EBRBS. Moreover, one of commonly used cluster computing, named Apache Spark, is then applied to implement the parallel rule generation and inference schemes of the Micro-EBRBS for big data multi- class classification problems. The comparative analyses of experimental studies demonstrate that the Micro-EBRBS not only can obtain a desired accuracy, but also has the comparatively better time complexity and computing efficiency than some popular classifiers, especially for multi-class classification problems.

AB - Big data classification problems have drawn great attention from diverse fields and many classifiers have been developed. Among those classifiers, the extended belief rule-based system (EBRBS) has shown its potential in both big data and multi-class situations while time complexity and computing efficiency are two challenging issues to be handled in EBRBS. As such, three improvements of EBRBS are proposed firstly in the present paper to decrease the time complexity and computing efficiency of EBRBS for multi-class classification under the assumption of large amount of data, including the strategy to skip rule weight calculation, a simplified evidential reasoning algorithm, and the domain division-based rule reduction method. This turns out to be a micro version of the EBRBS, classed Micro-EBRBS. Moreover, one of commonly used cluster computing, named Apache Spark, is then applied to implement the parallel rule generation and inference schemes of the Micro-EBRBS for big data multi- class classification problems. The comparative analyses of experimental studies demonstrate that the Micro-EBRBS not only can obtain a desired accuracy, but also has the comparatively better time complexity and computing efficiency than some popular classifiers, especially for multi-class classification problems.

U2 - 10.1109/TSMC.2018.2872843

DO - 10.1109/TSMC.2018.2872843

M3 - Article

SP - 1

EP - 21

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

T2 - IEEE Transactions on Systems, Man, and Cybernetics: Systems

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

SN - 2168-2216

ER -