Adaptive support vector machine for time-varying data streams using martingale

Authors:
Shen-Shyang Ho;Harry Wechsler
Affiliations:
George Mason University, Department of Computer Science, Fairfax, VA;George Mason University, Department of Computer Science, Fairfax, VA
Venue:
IJCAI'05 Proceedings of the 19th international joint conference on Artificial intelligence
Year:
2005

Citing 2
Cited 2

Mining time-changing data streams

Proceedings of the seventh ACM SIGKDD international conference on Knowledge discovery and data mining
Backpropagation applied to handwritten zip code recognition

Neural Computation

Detecting changes in unlabeled data streams using martingale

IJCAI'07 Proceedings of the 20th international joint conference on Artifical intelligence
Tracking a moving hypothesis for visual data with explicit switch detection

CISDA'09 Proceedings of the Second IEEE international conference on Computational intelligence for security and defense applications

Quantified Score

Hi-index	0.00

Visualization

Abstract

A martingale framework is proposed to enable support vector machine (SVM) to adapt to timevarying data streams. The adaptive SVM is a onepass incremental algorithm that (i) does not require a sliding window on the data stream, (ii) does not require monitoring the performance of the classifier as data points are streaming, and (iii) works well for high dimensional, multi-class data streams. Our experiments show that the novel adaptive SVM is effective at handling time-varying data streams simulated using both a synthetic dataset and a multiclass real dataset.