A model-selection framework for concept-drifting data streams

Bo Heng Chen, Kun Ta Chuang

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

There has been an increasing research interest in classification for data streams. Due to the evolving nature of data streams, it is a highly challenging issue to detect the appearance of concept drifts, which will make the current classification model invalid as time passes. So far most stream classification solutions exploit the so-called incremental learning process to continuously track the deviation of prediction accuracy. Unfortunately, to achieve the prompt concept-drifting detection, such strategies usually rely on an infeasible assumption about the availability of data instances with true labels. We in this paper propose a new framework, called Inference of Concept Evolution (abbreviated as ICE), to minimize the need of real-time acquisition of true labels. Specifically, the ICE framework is devised based on the idea of model reuse. The dictionary learning technique is utilized to determine whether the concept drift appears without the need of label acquisition. When the drift happens, the ICE framework will select the best model maintained in the model pool, decreasing the need of model re-training and its costly label acquisition. As demonstrated in our experimental result, the ICE framework can track the best model correctly and efficiently, showing its feasibility in real cases.

Original languageEnglish
Title of host publicationDSAA 2014 - Proceedings of the 2014 IEEE International Conference on Data Science and Advanced Analytics
EditorsGeorge Karypis, Longbing Cao, Wei Wang, Irwin King
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages290-296
Number of pages7
ISBN (Electronic)9781479969913
DOIs
Publication statusPublished - 2014 Mar 10
Event2014 IEEE International Conference on Data Science and Advanced Analytics, DSAA 2014 - Shanghai, China
Duration: 2014 Oct 302014 Nov 1

Publication series

NameDSAA 2014 - Proceedings of the 2014 IEEE International Conference on Data Science and Advanced Analytics

Other

Other2014 IEEE International Conference on Data Science and Advanced Analytics, DSAA 2014
Country/TerritoryChina
CityShanghai
Period14-10-3014-11-01

All Science Journal Classification (ASJC) codes

  • Artificial Intelligence
  • Information Systems
  • Information Systems and Management

Fingerprint

Dive into the research topics of 'A model-selection framework for concept-drifting data streams'. Together they form a unique fingerprint.

Cite this