Acoustic Footprints

Comparing Optimal Bounding Ellipsoid and Support Vector Machine Active Learning

TitleComparing Optimal Bounding Ellipsoid and Support Vector Machine Active Learning
Publication TypeConference Proceedings
Year of Conference2004
AuthorsGocken I, Joachim D, Deller JR
Conference Name17th International Conference on Pattern Recognition
Volume1
Pagination172-175
Date Published08/2004
Conference LocationCambridge UK
Abstract

In this paper we propose two active learning algorithms combining statistical active learning methods based on SVM and optimal bounding algorithms (OBE) of adaptive system identification. We unify SVM and OBE by demonstrating the similarities and representing SVM in an OBE interpretation. Samples are judiciously selected based on a volume measure provided by OBE using both simple heuristic and greedy optimal strategies. Preliminary experiments illustrate the effectiveness of the proposed algorithms as compared to similar methods.

URLhttp://doi.ieeecomputersociety.org/10.1109/ICPR.2004.1334041