An Empirical Study of Greedy Kernel Fisher Discriminants
Author
Source
Mathematical Problems in Engineering
Issue
Vol. 2015, Issue 2015 (31 Dec. 2015), pp.1-12, 12 p.
Publisher
Hindawi Publishing Corporation
Publication Date
2015-06-08
Country of Publication
Egypt
No. of Pages
12
Main Subjects
Abstract EN
A sparse version of Kernel Fisher Discriminant Analysis using an approach based on Matching Pursuit (MPKFDA) has been shown to be competitive with Kernel Fisher Discriminant Analysis and the Support Vector Machines on publicly available datasets, with additional experiments showing that MPKFDA on average outperforms these algorithms in extremely high dimensional settings.
In (nearly) all cases, the resulting classifier was sparser than the Support Vector Machine.
Natural questions that arise are what is the relative importance of the use of the Fisher criterion for selecting bases and the deflation step? Can we speed the algorithm up without degrading performance? Here we analyse the algorithm in more detail, providing alternatives to the optimisation criterion and the deflation procedure of the algorithm, and also propose a stagewise version.
We demonstrate empirically that these alternatives can provide considerable improvements in the computational complexity, whilst maintaining the performance of the original algorithm (and in some cases improving it).
American Psychological Association (APA)
Diethe, Tom. 2015. An Empirical Study of Greedy Kernel Fisher Discriminants. Mathematical Problems in Engineering،Vol. 2015, no. 2015, pp.1-12.
https://search.emarefa.net/detail/BIM-1074722
Modern Language Association (MLA)
Diethe, Tom. An Empirical Study of Greedy Kernel Fisher Discriminants. Mathematical Problems in Engineering No. 2015 (2015), pp.1-12.
https://search.emarefa.net/detail/BIM-1074722
American Medical Association (AMA)
Diethe, Tom. An Empirical Study of Greedy Kernel Fisher Discriminants. Mathematical Problems in Engineering. 2015. Vol. 2015, no. 2015, pp.1-12.
https://search.emarefa.net/detail/BIM-1074722
Data Type
Journal Articles
Language
English
Notes
Includes bibliographical references
Record ID
BIM-1074722