Advanced Query Strategies for Active Learning with Extreme Learning Machine
Abstract
This Work Proposes Three New Query Strategies for Active Learning. They Are Built on Modern Developments in Extreme Learning Machine (Elm): Class-Weighted Elm, Prediction Intervals with Elm, and Mislabeled Sam- Ples Detection with Elm. Both Elm and Active Learning Are Important Methods of Applied Machine Learning. Combined, They Offer a Fast and Precise Tool for Practical Data Acquisition in Classification Tasks Where Raw Data is Cheap But Labels Are Expensive to Get. Some Proposed Methods Rival the State-Of-The-Art in Performance and Speed, based on Testing with Three Real World Datasets.
Recommended Citation
A. Akusok et al., "Advanced Query Strategies for Active Learning with Extreme Learning Machine," ESANN 2017 - Proceedings, 25th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning, pp. 105 - 110, European Symposium on Artificial Neural Networks, Jan 2017.
Department(s)
Engineering Management and Systems Engineering
International Standard Book Number (ISBN)
978-287587039-1
Document Type
Article - Conference proceedings
Document Version
Citation
File Type
text
Language(s)
English
Rights
© 2024 European Symposium on Artificial Neural Networks, All rights reserved.
Publication Date
01 Jan 2017
