Kappa Updated Ensemble for drifting data stream mining
- PDF / 2,394,678 Bytes
- 44 Pages / 439.37 x 666.142 pts Page_size
- 67 Downloads / 208 Views
Kappa Updated Ensemble for drifting data stream mining Alberto Cano1
· Bartosz Krawczyk1
Received: 13 November 2018 / Revised: 29 June 2019 / Accepted: 6 September 2019 © The Author(s), under exclusive licence to Springer Science+Business Media LLC, part of Springer Nature 2019
Abstract Learning from data streams in the presence of concept drift is among the biggest challenges of contemporary machine learning. Algorithms designed for such scenarios must take into an account the potentially unbounded size of data, its constantly changing nature, and the requirement for real-time processing. Ensemble approaches for data stream mining have gained significant popularity, due to their high predictive capabilities and effective mechanisms for alleviating concept drift. In this paper, we propose a new ensemble method named Kappa Updated Ensemble (KUE). It is a combination of online and block-based ensemble approaches that uses Kappa statistic for dynamic weighting and selection of base classifiers. In order to achieve a higher diversity among base learners, each of them is trained using a different subset of features and updated with new instances with given probability following a Poisson distribution. Furthermore, we update the ensemble with new classifiers only when they contribute positively to the improvement of the quality of the ensemble. Finally, each base classifier in KUE is capable of abstaining itself for taking a part in voting, thus increasing the overall robustness of KUE. An extensive experimental study shows that KUE is capable of outperforming state-of-the-art ensembles on standard and imbalanced drifting data streams while having a low computational complexity. Moreover, we analyze the use of Kappa versus accuracy to drive the criterion to select and update the classifiers, the contribution of the abstaining mechanism, the contribution of the diversification of classifiers, and the contribution of the hybrid architecture to update the classifiers in an online manner. Keywords Machine learning · Data streams · Concept drift · Classification · Ensemble learning
1 Introduction The data revolution over the last two decades has changed almost every aspect of data analytics. One must take into account the fact that the size of data is constantly growing and
Editor: João Gama.
B
Alberto Cano [email protected] Bartosz Krawczyk [email protected]
1
Virginia Commonwealth University, 401 W. Main St. E4251, Richmond, VA 23284, USA
123
Machine Learning
one cannot store all of it. Data is in motion, constantly expanding, and changing its properties (Morales et al. 2016). Additionally, data may come from many sources at the same time, calling for efficient preprocessing and standardization (Ramirez-Gallego et al. 2017). Such changes affected various real-life applications, including social media (Miller et al. 2014), medicine (Triantafyllopoulos et al. 2016), and security (Faisal et al. 2015) to name a few. This poses challenges for learning systems that must accommodate all these properties, while maintaining
Data Loading...
