Title :
Active learning for reducing bias and variance of a classifier using Jensen-Shannon divergence
Author_Institution :
Dept. of Comput. Sci., State Univ. of New York, Albany, NY, USA
Abstract :
We consider reducing loss of a classifier by decreasing its bias and variance. Embarking upon classification of scarcely labeled data, we use active learning approach in semi-supervised learning, and show that we can speed up convergence to a desired level of loss. Our focus, in this paper, is on the best instance selection for labeling the unlabeled data; we use Jensen-Shannon divergence as one selection criterion. We show that our single instance selection approaches are superior to multiple selection approach. Empirical results indicate that this method can decrease classification loss significantly.
Keywords :
learning (artificial intelligence); pattern classification; Jensen-Shannon divergence; active learning; best instance selection; data classification; data labeling; multiple selection approach; semisupervised learning; single instance selection; Bagging; Bayesian methods; Computer science; Convergence; Humans; Labeling; Learning systems; Machine learning; Monte Carlo methods; Semisupervised learning;
Conference_Titel :
Machine Learning and Applications, 2005. Proceedings. Fourth International Conference on
Print_ISBN :
0-7695-2495-8
DOI :
10.1109/ICMLA.2005.7