A comparison of multiple instance and group based learning

Brossi, Steven D. and Bradley, Andrew P. (2012). A comparison of multiple instance and group based learning. In: Geoff West and Peter Kovesi, 14th International Conference on Digital Image Computing Techniques and Applications, DICTA 2012: Proceedings. 14th International Conference on Digital Image Computing Techniques and Applications, DICTA 2012, Fremantle, WA, Australia, (1-8). 3 - 5 December 2012. doi:10.1109/DICTA.2012.6411737

Attached Files (Some files may be inaccessible until you login with your UQ eSpace credentials)
Name Description MIMEType Size Downloads

Author Brossi, Steven D.
Bradley, Andrew P.
Title of paper A comparison of multiple instance and group based learning
Conference name 14th International Conference on Digital Image Computing Techniques and Applications, DICTA 2012
Conference location Fremantle, WA, Australia
Conference dates 3 - 5 December 2012
Proceedings title 14th International Conference on Digital Image Computing Techniques and Applications, DICTA 2012: Proceedings
Journal name 2012 International Conference On Digital Image Computing Techniques and Applications (Dicta)
Place of Publication Piscataway, NJ, United States
Publisher IEEE
Publication Year 2012
Sub-type Fully published paper
DOI 10.1109/DICTA.2012.6411737
ISBN 9781467321815
Editor Geoff West
Peter Kovesi
Start page 1
End page 8
Total pages 8
Collection year 2013
Language eng
Abstract/Summary In this paper we compare the performance of a number of multiple-instance learning (MIL) and group based (GB) classification algorithms on both a synthetic and real-world Pap smear dataset. We utilise the synthetic dataset to demonstrate that performance improves as both bag size and percent positives increase and that MIL outperforms GB algorithms when the percentage positives is less than 50%. However, as the positive bags become increasingly homogeneous, as is apparent on the real-world dataset, the two approaches become comparable. This result highlights that the performance of a MIL or GB algorithm will be maximised when the algorithm's MIL assumption matches the reality of the dataset. Therefore, on the Pap smear dataset, algorithms with a more generalised MIL assumption demonstrate the strongest performance.
Q-Index Code E1
Q-Index Status Confirmed Code
Institutional Status UQ

 
Versions
Version Filter Type
Citation counts: TR Web of Science Citation Count  Cited 0 times in Thomson Reuters Web of Science Article
Scopus Citation Count Cited 0 times in Scopus Article
Google Scholar Search Google Scholar
Created: Wed, 03 Apr 2013, 14:23:52 EST by Andrew Bradley on behalf of School of Information Technol and Elec Engineering