Hellinger distance based drift detection for nonstationary environments
Authors:
- 1 April 2011
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 41-48
Abstract
Most machine learning algorithms, including many online learners, assume that the data distribution to be learned is fixed. There are many real-world problems where the distribution of the data changes as a function of time. Changes in nonstationary data distributions can significantly reduce the generalization ability of the learning algorithm on new or field data, if the algorithm is not equipped to track such changes. When the stationary data distribution assumption does not hold, the learner must take appropriate actions to ensure that the new/relevant information is learned. On the other hand, data distributions do not necessarily change continuously, necessitating the ability to monitor the distribution and detect when a significant change in distribution has occurred. In this work, we propose and analyze a feature based drift detection method using the Hellinger distance to detect gradual or abrupt changes in the distribution.Keywords
This publication has 12 references indexed in Scilit:
- Change detection tests using the ICI rule2010
- Just in time classifiers: Managing the slow drift case2009
- Just-in-Time Adaptive Classifiers—Part II: Designing the ClassifierIEEE Transactions on Neural Networks, 2008
- Just-in-Time Adaptive Classifiers—Part I: Detecting Nonstationary ChangesIEEE Transactions on Neural Networks, 2008
- A framework for monitoring classifiers’ performance: when and why failure occurs?Knowledge and Information Systems, 2008
- An Ensemble Approach for Incremental Learning in Nonstationary Environments2007
- Learn++: an incremental learning algorithm for supervised neural networksIEEE Transactions on Systems, Man and Cybernetics, Part C (Applications and Reviews), 2001
- A streaming ensemble algorithm (SEA) for large-scale classification2001
- Life-long learning Cell Structures—continuously learning without catastrophic interferenceNeural Networks, 2001
- A cumulative sum type of method for environmental monitoringEnvironmetrics, 2000