An adaptive ensemble classifier for mining concept drifting data streams

Dewan Md. Farid, Li Zhang, Alamgir Hossain, Chowdhury Mofizur Rahman, Rebecca Strachan, Graham Sexton, Keshav Dahal

Research output: Contribution to journalArticle

Abstract

It is challenging to use traditional data mining techniques to deal with real-time data stream classifications. Existing mining classifiers need to be updated frequently to adapt to the changes in data streams. To address this issue, in this paper we propose an adaptive ensemble approach for classification and novel class detection in concept drifting data streams. The proposed approach uses traditional mining classifiers and updates the ensemble model automatically so that it represents the most recent concepts in data streams. For novel class detection we consider the idea that data points belonging to the same class should be closer to each other and should be far apart from the data points belonging to other classes. If a data point is well separated from the existing data clusters, it is identified as a novel class instance. We tested the performance of this proposed stream classification model against that of existing mining algorithms using real benchmark datasets from UCI (University of California, Irvine) machine learning repository. The experimental results prove that our approach shows great flexibility and robustness in novel class detection in concept drifting and outperforms traditional classification models in challenging real-life data stream applications.
Original languageEnglish
Pages (from-to)5895-5906
JournalExpert Systems with Applications
Volume40
Issue number15
DOIs
Publication statusPublished - 1 Nov 2013
Externally publishedYes

Keywords

  • Adaptive ensembles
  • Concept drift
  • Clustering
  • Data streams
  • Decision trees
  • Novel classes

Cite this

Farid, D. M., Zhang, L., Hossain, A., Rahman, C. M., Strachan, R., Sexton, G., & Dahal, K. (2013). An adaptive ensemble classifier for mining concept drifting data streams. Expert Systems with Applications, 40(15), 5895-5906. https://doi.org/10.1016/j.eswa.2013.05.001
Farid, Dewan Md. ; Zhang, Li ; Hossain, Alamgir ; Rahman, Chowdhury Mofizur ; Strachan, Rebecca ; Sexton, Graham ; Dahal, Keshav. / An adaptive ensemble classifier for mining concept drifting data streams. In: Expert Systems with Applications. 2013 ; Vol. 40, No. 15. pp. 5895-5906.
@article{a3ed1707efd34e75a8405b349123c06b,
title = "An adaptive ensemble classifier for mining concept drifting data streams",
abstract = "It is challenging to use traditional data mining techniques to deal with real-time data stream classifications. Existing mining classifiers need to be updated frequently to adapt to the changes in data streams. To address this issue, in this paper we propose an adaptive ensemble approach for classification and novel class detection in concept drifting data streams. The proposed approach uses traditional mining classifiers and updates the ensemble model automatically so that it represents the most recent concepts in data streams. For novel class detection we consider the idea that data points belonging to the same class should be closer to each other and should be far apart from the data points belonging to other classes. If a data point is well separated from the existing data clusters, it is identified as a novel class instance. We tested the performance of this proposed stream classification model against that of existing mining algorithms using real benchmark datasets from UCI (University of California, Irvine) machine learning repository. The experimental results prove that our approach shows great flexibility and robustness in novel class detection in concept drifting and outperforms traditional classification models in challenging real-life data stream applications.",
keywords = "Adaptive ensembles, Concept drift, Clustering, Data streams, Decision trees, Novel classes",
author = "Farid, {Dewan Md.} and Li Zhang and Alamgir Hossain and Rahman, {Chowdhury Mofizur} and Rebecca Strachan and Graham Sexton and Keshav Dahal",
year = "2013",
month = "11",
day = "1",
doi = "10.1016/j.eswa.2013.05.001",
language = "English",
volume = "40",
pages = "5895--5906",
journal = "Expert Systems with Applications",
issn = "0957-4174",
publisher = "Elsevier B.V.",
number = "15",

}

Farid, DM, Zhang, L, Hossain, A, Rahman, CM, Strachan, R, Sexton, G & Dahal, K 2013, 'An adaptive ensemble classifier for mining concept drifting data streams' Expert Systems with Applications, vol. 40, no. 15, pp. 5895-5906. https://doi.org/10.1016/j.eswa.2013.05.001

An adaptive ensemble classifier for mining concept drifting data streams. / Farid, Dewan Md.; Zhang, Li; Hossain, Alamgir; Rahman, Chowdhury Mofizur; Strachan, Rebecca; Sexton, Graham; Dahal, Keshav.

In: Expert Systems with Applications, Vol. 40, No. 15, 01.11.2013, p. 5895-5906.

Research output: Contribution to journalArticle

TY - JOUR

T1 - An adaptive ensemble classifier for mining concept drifting data streams

AU - Farid, Dewan Md.

AU - Zhang, Li

AU - Hossain, Alamgir

AU - Rahman, Chowdhury Mofizur

AU - Strachan, Rebecca

AU - Sexton, Graham

AU - Dahal, Keshav

PY - 2013/11/1

Y1 - 2013/11/1

N2 - It is challenging to use traditional data mining techniques to deal with real-time data stream classifications. Existing mining classifiers need to be updated frequently to adapt to the changes in data streams. To address this issue, in this paper we propose an adaptive ensemble approach for classification and novel class detection in concept drifting data streams. The proposed approach uses traditional mining classifiers and updates the ensemble model automatically so that it represents the most recent concepts in data streams. For novel class detection we consider the idea that data points belonging to the same class should be closer to each other and should be far apart from the data points belonging to other classes. If a data point is well separated from the existing data clusters, it is identified as a novel class instance. We tested the performance of this proposed stream classification model against that of existing mining algorithms using real benchmark datasets from UCI (University of California, Irvine) machine learning repository. The experimental results prove that our approach shows great flexibility and robustness in novel class detection in concept drifting and outperforms traditional classification models in challenging real-life data stream applications.

AB - It is challenging to use traditional data mining techniques to deal with real-time data stream classifications. Existing mining classifiers need to be updated frequently to adapt to the changes in data streams. To address this issue, in this paper we propose an adaptive ensemble approach for classification and novel class detection in concept drifting data streams. The proposed approach uses traditional mining classifiers and updates the ensemble model automatically so that it represents the most recent concepts in data streams. For novel class detection we consider the idea that data points belonging to the same class should be closer to each other and should be far apart from the data points belonging to other classes. If a data point is well separated from the existing data clusters, it is identified as a novel class instance. We tested the performance of this proposed stream classification model against that of existing mining algorithms using real benchmark datasets from UCI (University of California, Irvine) machine learning repository. The experimental results prove that our approach shows great flexibility and robustness in novel class detection in concept drifting and outperforms traditional classification models in challenging real-life data stream applications.

KW - Adaptive ensembles

KW - Concept drift

KW - Clustering

KW - Data streams

KW - Decision trees

KW - Novel classes

U2 - 10.1016/j.eswa.2013.05.001

DO - 10.1016/j.eswa.2013.05.001

M3 - Article

VL - 40

SP - 5895

EP - 5906

JO - Expert Systems with Applications

JF - Expert Systems with Applications

SN - 0957-4174

IS - 15

ER -

Farid DM, Zhang L, Hossain A, Rahman CM, Strachan R, Sexton G et al. An adaptive ensemble classifier for mining concept drifting data streams. Expert Systems with Applications. 2013 Nov 1;40(15):5895-5906. https://doi.org/10.1016/j.eswa.2013.05.001