A taxonomic look at instance-based stream classifier

Allbwn ymchwil: Cyfraniad at gyfnodolynErthygladolygiad gan gymheiriaid

StandardStandard

A taxonomic look at instance-based stream classifier. / Gunn, Iain; Arnaiz-Gonzalez, Alvar; Kuncheva, Ludmila.
Yn: Neurocomputing, Cyfrol 286, 19.04.2018, t. 167–178.

Allbwn ymchwil: Cyfraniad at gyfnodolynErthygladolygiad gan gymheiriaid

HarvardHarvard

APA

CBE

Gunn I, Arnaiz-Gonzalez A, Kuncheva L. 2018. A taxonomic look at instance-based stream classifier. Neurocomputing. 286:167–178.

MLA

Gunn, Iain, Alvar Arnaiz-Gonzalez, a Ludmila Kuncheva. "A taxonomic look at instance-based stream classifier". Neurocomputing. 2018, 286. 167–178.

VancouverVancouver

Gunn I, Arnaiz-Gonzalez A, Kuncheva L. A taxonomic look at instance-based stream classifier. Neurocomputing. 2018 Ebr 19;286:167–178. Epub 2018 Chw 2.

Author

Gunn, Iain ; Arnaiz-Gonzalez, Alvar ; Kuncheva, Ludmila. / A taxonomic look at instance-based stream classifier. Yn: Neurocomputing. 2018 ; Cyfrol 286. tt. 167–178.

RIS

TY - JOUR

T1 - A taxonomic look at instance-based stream classifier

AU - Gunn, Iain

AU - Arnaiz-Gonzalez, Alvar

AU - Kuncheva, Ludmila

PY - 2018/4/19

Y1 - 2018/4/19

N2 - Large numbers of data streams are today generated in many fields. A key challenge when learning from such streams is the problem of concept drift. Many methods, including many prototype methods, have been proposed in recent years to address this problem. This paper presents a refined taxonomy of instance selection and generation methods for the classification of data streams subject to concept drift. The taxonomy allows discrimination among a large number of methods which pre-existing taxonomies for offline instance selection methods did not distinguish. This makes possible a valuable new perspective on experimental results, and provides a framework for discussion of the concepts behind different algorithm-design approaches. We review a selection of modern algorithms for the purpose of illustrating the distinctions made by the taxonomy. We present the results of a numerical experiment which examined the performance of a number of representative methods on both synthetic and real -world data setswith and without concept drift, and discuss the implications for the directions of future research in lightof the taxonomy. On the basis of the experimental results, we are able to give recommendations for theexperimental evaluation of algorithms which may be proposed in the future.

AB - Large numbers of data streams are today generated in many fields. A key challenge when learning from such streams is the problem of concept drift. Many methods, including many prototype methods, have been proposed in recent years to address this problem. This paper presents a refined taxonomy of instance selection and generation methods for the classification of data streams subject to concept drift. The taxonomy allows discrimination among a large number of methods which pre-existing taxonomies for offline instance selection methods did not distinguish. This makes possible a valuable new perspective on experimental results, and provides a framework for discussion of the concepts behind different algorithm-design approaches. We review a selection of modern algorithms for the purpose of illustrating the distinctions made by the taxonomy. We present the results of a numerical experiment which examined the performance of a number of representative methods on both synthetic and real -world data setswith and without concept drift, and discuss the implications for the directions of future research in lightof the taxonomy. On the basis of the experimental results, we are able to give recommendations for theexperimental evaluation of algorithms which may be proposed in the future.

M3 - Article

VL - 286

SP - 167

EP - 178

JO - Neurocomputing

JF - Neurocomputing

SN - 0925-2312

ER -