Abstract
Classification systems meant to operate in non-stationary environments are requested to adapt when the process generating the observed data changes. A particularly effective form of adaptation in the abrupt perturbation case suggests to release the obsolete knowledge base of the classifier (or training set), and consider novel samples to configure the new classification model. In this direction, we propose an adaptive classifier based on a change detection test used both for detecting changes in the process and identifying the new training set (and, then, the new classifier). A key point of the proposed solution is that no assumptions are made about the distribution of the process generating the data. Experimental results show that the proposed adaptive classification system is particularly effective in situations where the process generating the data evolves through a sequence of abrupt changes.
This is a preview of subscription content, log in via an institution.
Buying options
Tax calculation will be finalised at checkout
Purchases are for personal use only
Learn about institutional subscriptionsPreview
Unable to display preview. Download preview PDF.
References
Helmbold, D.P., Long, P.M.: Tracking drifting concepts by minimizing disagreements. Machine Learning 14(1), 27–45 (1994)
Kuh, A., Petsche, T.: Learning Time Varying Concepts with Applications to Pattern Recognition Problems. In: Petsche, T. (ed.) Conference Record Twenty-Fourth Asilomar Conference on Signals, Systems and Computers, 1990, vol. 2, p. 971 (1990)
Tsymbal. A.: The problem of concept drift: definitions and related work. Technical Report: Trinity College, Dublin, Ireland, TCD-CS-2004-15 (2004)
Widmer, G., Kubat, M.: Learning in the presence of concept drift and hidden contexts. Machine Learning 23(1), 69–101 (1996)
Klinkenberg, R.: Learning drifting concepts: example selection vs. example weighting, Intelligent Data Analysis. Special Issue on Incremental Learning Systems Capable of Dealing with Concept Drift 8(3) (2004)
Alippi, C., Roveri, M.: Just-in-time adaptive classifiers—Part I: Detecting nonstationarity changes. IEEE Transactions on Neural Networks 19 (2008)
Alippi, C., Roveri, M.: Just-in-Time Adaptive Classifiers–Part II: Designing the classifier. IEEE Transactions on Neural Networks 19(11), 2053–2064 (2008)
Alippi, C., Boracchi, G., Roveri, M.: Just in time classifiers: Managing the slow drift case. In: Proc. of the IEEE 2009 International Joint Conference on Neural Networks 2009, June 14-19, pp. 114–120 (2009)
Alippi, C., Boracchi, G., Roveri, M.: Change Detection Tests Using the ICI rule. In: Proc. of the IEEE 2010 International Joint Conference on Neural Networks, June 18-23 (2010)
Fukunaga, K.: Introduction to Statistical Pattern Recognition. Academic, New York (1972)
Goldenshluger, A., Nemirovski, A.: On spatial adaptive estimation of nonparametric regression. Mathematical Methods of Statistics 6, 135–170 (1997)
Katkovnik, V.: A new method for varying adaptive bandwidth selection. IEEE Transactions. on Signal Processing 47(9), 2567–2571 (1999)
Mudholkar, G.S., Trivedi, M.C.: A Gaussian Approximation to the Distribution of the Sample Variance for Nonnormal Populations. Journal of the American Statistical Association 76(374), 479–485 (1981)
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2010 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Alippi, C., Boracchi, G., Roveri, M. (2010). Adaptive Classifiers with ICI-Based Adaptive Knowledge Base Management. In: Diamantaras, K., Duch, W., Iliadis, L.S. (eds) Artificial Neural Networks – ICANN 2010. ICANN 2010. Lecture Notes in Computer Science, vol 6353. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-642-15822-3_56
Download citation
DOI: https://doi.org/10.1007/978-3-642-15822-3_56
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-642-15821-6
Online ISBN: 978-3-642-15822-3
eBook Packages: Computer ScienceComputer Science (R0)