Batch-Incremental versus Instance-Incremental Learning in Dynamic and Evolving Data
Abstract
Many real world problems involve the challenging context of data streams, where classifiers must be incremental: able to learn from a theoretically-infinite stream of examples using limited time and memory, while being able to predict at any point. Two approaches dominate the literature: batch-incremental methods that gather examples in batches to train models; and instance-incremental methods that learn from each example as it arrives. Typically, papers in the literature choose one of these approaches, but provide insufficient evidence or references to justify their choice. We provide a first in-depth analysis comparing both approaches, including how they adapt to concept drift, and an extensive empirical study to compare several different versions of each approach. Our results reveal the respective advantages and disadvantages of the methods, which we discuss in detail.
Keywords
data streams incremental dynamic evolving on-linePreview
Unable to display preview. Download preview PDF.
References
- 1.Bifet, A., Holmes, G., Pfahringer, B., Kirkby, R., Gavaldà, R.: New ensemble methods for evolving data streams. In: KDD, pp. 139–148 (2009)Google Scholar
- 2.Beringer, J., Hüllermeier, E.: Efficient instance-based learning on data streams. Intelligent Data Analysis 11(6), 627–650 (2007)Google Scholar
- 3.Zhang, P., Gao, B.J., Zhu, X., Guo, L.: Enabling fast lazy learning for data streams. In: ICDM, pp. 932–941 (2011)Google Scholar
- 4.John, G.H., Langley, P.: Estimating continuous distributions in bayesian classifiers. In: Eleventh Conference on Uncertainty in Artificial Intelligence, San Mateo, pp. 338–345. Morgan Kaufmann (1995)Google Scholar
- 5.Domingos, P., Hulten, G.: Mining high-speed data streams. In: KDD, pp. 71–80 (2000)Google Scholar
- 6.Bifet, A., Gavaldà, R.: Learning from time-changing data with adaptive windowing. In: SDM (2007)Google Scholar
- 7.Holte, R.C.: Very simple classification rules perform well on most commonly used datasets. Machine Learning 11, 63–91 (1993)zbMATHCrossRefGoogle Scholar
- 8.Caruana, R., Niculescu-Mizil, A.: An empirical comparison of supervised learning algorithms. In: ICML, pp. 161–168 (2006)Google Scholar
- 9.Bottou, L.: Online algorithms and stochastic approximations. Online Learning and Neural Networks (1998)Google Scholar
- 10.Oza, N.C., Russell, S.J.: Experimental comparisons of online and batch versions of bagging and boosting. In: KDD, pp. 359–364 (2001)Google Scholar
- 11.Oza, N., Russell, S.: Online bagging and boosting. In: Artificial Intelligence and Statistics 2001, pp. 105–112. Morgan Kaufmann (2001)Google Scholar
- 12.Bifet, A., Gavaldà, R.: Adaptive Learning from Evolving Data Streams. In: Adams, N.M., Robardet, C., Siebes, A., Boulicaut, J.-F. (eds.) IDA 2009. LNCS, vol. 5772, pp. 249–260. Springer, Heidelberg (2009)CrossRefGoogle Scholar
- 13.Bifet, A., Holmes, G., Pfahringer, B.: Leveraging Bagging for Evolving Data Streams. In: Balcázar, J.L., Bonchi, F., Gionis, A., Sebag, M. (eds.) ECML PKDD 2010, Part I. LNCS, vol. 6321, pp. 135–150. Springer, Heidelberg (2010)CrossRefGoogle Scholar
- 14.Qu, W., Zhang, Y., Zhu, J., Qiu, Q.: Mining Multi-label Concept-Drifting Data Streams Using Dynamic Classifier Ensemble. In: Zhou, Z.-H., Washio, T. (eds.) ACML 2009. LNCS, vol. 5828, pp. 308–321. Springer, Heidelberg (2009)CrossRefGoogle Scholar
- 15.Wang, H., Fan, W., Yu, P.S., Han, J.: Mining concept-drifting data streams using ensemble classifiers. In: KDD 2003, pp. 226–235. ACM, New York (2003)CrossRefGoogle Scholar
- 16.Spyromitros-Xioufis, E., Spiliopoulou, M., Tsoumakas, G., Vlahavas, I.: Dealing with concept drift and class imbalance in multi-label stream classification. In: IJCAI, pp. 1583–1588 (2011)Google Scholar
- 17.Bifet, A., Holmes, G., Kirkby, R., Pfahringer, B.: MOA: Massive Online Analysis. Journal of Machine Learning Research, JMLR (2010)Google Scholar
- 18.Street, W.N., Kim, Y.: A streaming ensemble algorithm (SEA) for large-scale classification. In: KDD, pp. 377–382 (2001)Google Scholar
- 19.Hulten, G., Spencer, L., Domingos, P.: Mining time-changing data streams. In: KDD, pp. 97–106 (2001)Google Scholar
- 20.Breiman, L., Friedman, J.H., Olshen, R.A., Stone, C.J.: Classification and Regression Trees. Wadsworth (1984)Google Scholar
- 21.Asuncion, A., Newman, D.: UCI machine learning repository (2007)Google Scholar
- 22.Gama, J., Rocha, R., Medas, P.: Accurate decision trees for mining high-speed data streams. In: KDD, pp. 523–528 (2003)Google Scholar
- 23.Harries, M.: Splice-2 comparative evaluation: Electricity pricing. Technical report, The University of South Wales (1999)Google Scholar
- 24.Gama, J., Medas, P., Castillo, G., Rodrigues, P.: Learning with Drift Detection. In: Bazzan, A.L.C., Labidi, S. (eds.) SBIA 2004. LNCS (LNAI), vol. 3171, pp. 286–295. Springer, Heidelberg (2004)CrossRefGoogle Scholar
- 25.Lang, K.: The 20 newsgroups dataset (2008), http://people.csail.mit.edu/jrennie/20Newsgroups/
- 26.Read, J., Bifet, A., Holmes, G., Pfahringer, B.: Scalable and efficient multi-label classification for evolving data streams. Machine Learning, 1–30 (2012)Google Scholar
- 27.Demšar, J.: Statistical comparisons of classifiers over multiple data sets. The Journal of Machine Learning Research 7, 1–30 (2006)zbMATHGoogle Scholar
- 28.Bifet, A., Holmes, G., Pfahringer, B., Frank, E.: Fast Perceptron Decision Tree Learning from Evolving Data Streams. In: Zaki, M.J., Yu, J.X., Ravindran, B., Pudi, V. (eds.) PAKDD 2010. LNCS, vol. 6119, pp. 299–310. Springer, Heidelberg (2010)CrossRefGoogle Scholar