Batch weighted ensemble for mining data streams with concept drift

Authors:
Magdalena Deckert
Affiliations:
Institute of Computing Science, Poznań University of Technology, Poznań, Poland
Venue:
ISMIS'11 Proceedings of the 19th international conference on Foundations of intelligent systems
Year:
2011

Citing 5
Cited 1

Learning in the presence of concept drift and hidden contexts

Machine Learning
Mining concept-drifting data streams using ensemble classifiers

Proceedings of the ninth ACM SIGKDD international conference on Knowledge discovery and data mining
Mining Data Streams with Labeled and Unlabeled Training Examples

ICDM '09 Proceedings of the 2009 Ninth IEEE International Conference on Data Mining
Knowledge Discovery from Data Streams

Knowledge Discovery from Data Streams
MOA: Massive Online Analysis

The Journal of Machine Learning Research

Combining block-based and online methods in learning ensembles from concept drifting data streams

Information Sciences: an International Journal

Quantified Score

Hi-index	0.00

Visualization

Abstract

This paper presents a new framework for dealing with two main types of concept drift: sudden and gradual drift in labelled data with decision attribute. The learning examples are processed in batches of the same size. This new framework, called Batch Weighted Ensemble, is based on incorporating drift detector into the evolving ensemble. Its performance was evaluated experimentaly on data sets with different types of concept drift and compared with the performance of a standard Accuracy Weighted Ensemble classifier. The results show that BWE improves evaluation measures like processing time, memory used and obtain competitive total accuracy.