Adaptive mechanisms for classification problems with drifting data

Authors: Sahel, Z., Bouchachia, A., Gabrys, B. and Rogers, P.

Journal: Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

Volume: 4693 LNAI

Issue: PART 2

Pages: 419-426

eISSN: 1611-3349

ISBN: 9783540748267

ISSN: 0302-9743

DOI: 10.1007/978-3-540-74827-4_53

Abstract:

Most work on supervised learning is undertaken on static problems. However, in many real world classification problems, the environment in which the classifiers operate is dynamic i.e. the descriptions of classes change with time. In this paper, the process of generating drifting data is introduced in order to assess two adaptive approaches that deal with dynamically changing data. These approaches are: retraining on evolving data set and incremental learning. The empirical evaluation has shown that both these approaches improve the performance compared to the non-adaptive mode though a number of outstanding research issues remain. © Springer-Verlag Berlin Heidelberg 2007.

Source: Scopus