TY - GEN
T1 - Online learning by ellipsoid method
AU - Yang, Liu
AU - Jin, Rong
AU - Ye, Jieping
PY - 2009
Y1 - 2009
N2 - In this work, we extend the ellipsoid method, which was originally designed for convex optimization, for online learning. The key idea is to approximate by an ellipsoid the classification hypotheses that are consistent with all the training examples received so far. This is in contrast to most online learning algorithms where only a single classifier is maintained at each iteration. Efficient algorithms are presented for updating both the centroid and the positive definite matrix of ellipsoid given a misclassified example. In addition to the classical ellipsoid method, an improved version for online learning is also presented. Mistake bounds for both ellipsoid methods are derived. Evaluation with the USPS dataset and three UCI data-sets shows encouraging results when comparing the proposed online learning algorithm to two state-of-the-art online learners.
AB - In this work, we extend the ellipsoid method, which was originally designed for convex optimization, for online learning. The key idea is to approximate by an ellipsoid the classification hypotheses that are consistent with all the training examples received so far. This is in contrast to most online learning algorithms where only a single classifier is maintained at each iteration. Efficient algorithms are presented for updating both the centroid and the positive definite matrix of ellipsoid given a misclassified example. In addition to the classical ellipsoid method, an improved version for online learning is also presented. Mistake bounds for both ellipsoid methods are derived. Evaluation with the USPS dataset and three UCI data-sets shows encouraging results when comparing the proposed online learning algorithm to two state-of-the-art online learners.
UR - http://www.scopus.com/inward/record.url?scp=70049083822&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=70049083822&partnerID=8YFLogxK
U2 - 10.1145/1553374.1553521
DO - 10.1145/1553374.1553521
M3 - Conference contribution
AN - SCOPUS:70049083822
SN - 9781605585161
T3 - ACM International Conference Proceeding Series
BT - Proceedings of the 26th Annual International Conference on Machine Learning, ICML'09
T2 - 26th Annual International Conference on Machine Learning, ICML'09
Y2 - 14 June 2009 through 18 June 2009
ER -