Active Learning With Unsupervised Ensembles Of Classifiers
Panagiotis Traganitis, Dimitrios Berberidis, Georgios B. Giannakis
-
Members: FreeSPS
IEEE Members: $11.00
Non-members: $15.00Length: 15:30
04 May 2020
The present work introduces a simple scheme for active classification of data using unsupervised ensembles of classifiers. Uncertainty sampling, with different uncertainty measures, is evaluated for data selection, while an online expectation maximization algorithm is derived to estimate model parameters on-the-fly. Preliminary tests on real data showcase the potential of the novel approach.
