Adaptive tangent distance classifier on recognition of handwritten digits

Shuen-Lin Jeng and Yu-Te Liu

Journal of Applied Statistics, 2011, vol. 38, issue 11, 2647-2659

Abstract: Simard et al. [1617] proposed a transformation distance called “tangent distance” (TD) which can make pattern recognition be efficient. The key idea is to construct a distance measure which is invariant with respect to some chosen transformations. In this research, we provide a method using adaptive TD based on an idea inspired by “discriminant adaptive nearest neighbor” [7]. This method is relatively easy compared with many other complicated ones. A real handwritten recognition data set is used to illustrate our new method. Our results demonstrate that the proposed method gives lower classification error rates than those by standard implementation of neural networks and support vector machines and is as good as several other complicated approaches.

Date: 2011
References: View complete reference list from CitEc
Citations:

Downloads: (external link)
http://hdl.handle.net/10.1080/02664763.2011.567247 (text/html)
Access to full text is restricted to subscribers.

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:taf:japsta:v:38:y:2011:i:11:p:2647-2659

Ordering information: This journal article can be ordered from
http://www.tandfonline.com/pricing/journal/CJAS20

DOI: 10.1080/02664763.2011.567247

Access Statistics for this article

Journal of Applied Statistics is currently edited by Robert Aykroyd

More articles in Journal of Applied Statistics from Taylor & Francis Journals
Bibliographic data for series maintained by Chris Longhurst ().