Abstract—With the emergence of large-volume and high speed streaming data, mining of stream data has become a focus on increasing interests in real applications including credit card fraud detection, target marketing, network intrusion detection, etc. The major new challenges in stream data mining are, (a) since streaming data may flow in and out indefinitely and in fast speed, it is usually expected that a stream data mining process can only scan a data once and (b) since the characteristics of the data may evolve over time, it is desirable to incorporate evolving features of streaming data. This paper introduced new adaptive ensemble boosting approach for the classification of streaming data with concept drift. This adaptive ensemble boosting method uses adaptive sliding window and Hoeffding Tree with naïve bayes adaptive as base learner. The result shows that the proposed algorithm works well in changing environment as compared with other ensemble classifiers.
Index Terms—Concept drift, ensemble approach, hoeffding tree, sliding window, stream data.
K. K. Wankhade was with Priyadarshini College of Engineering, Nagpur INDIA 440019. He is now Assistant Professor with the Department of Information Technolgy, G. H. Raisoni College of Engineering, Nagpur, Maharashtra, INDIA 440019 (firstname.lastname@example.org).
S. S. Dongre was with Sinhgad Institute of Technology, Lonavala, Pune, INDIA. She is now Assistant Professor with the Department of Computer Science and Engineering, G. H. Raisoni College of Engineering, Nagpur, Maharashtra, INDIA 440019 (email@example.com).
Cite: Kapil K. Wankhade and Snehlata S. Dongre, "A New Adaptive Ensemble Boosting Classifier for Concept Drifting Stream Data," International Journal of Modeling and Optimization vol. 2, no. 4, pp. 493-497, 2012.