Dissertation/Thesis

A Windowed Version Of The Bunke And Bühler Algorithm To Better Approximate Dissimilarities Between Cyclic Patterns.

Bibliographic Details
Title: A Windowed Version Of The Bunke And Bühler Algorithm To Better Approximate Dissimilarities Between Cyclic Patterns.
Authors: R. A. Mollineda, E. Vidal, F. Casacuberta
Contributors: The Pennsylvania State University CiteSeerX Archives
Source: http://www.iti.upv.es/~prhlt/PAPERS/papers96-00/2000/Mollineda00e.ps.gz.
Publication Year: 2000
Collection: CiteSeerX
Subject Terms: Cyclic patterns, cyclic strings, approximate string matching, structural pattern analysis, 2D shape recognition
Description: An ecient approximate technique for measuring dissimilarities between cyclic patterns is presented. It is inspired on the quadratic time algorithm proposed by Bunke and Buhler (BBA). This algorithm computes an estimation of the distance measured between two cyclic strings x and y, by nding in an edit graph, an alignment between x with its most similar substring of some rotation of y. The technique presented in this paper de nes a diagonal window on the edit graph for each starting node where alignments originate from. Each alignment will be constrained to be within the window associated to its starting node. It guarantees alignments between x and substrings very close to some rotation of y, leading to a much better approximation to the exact cyclic distance value between x and y. Experiments were conducted on both arti cial and real data, to demonstrate the capabilities of the new technique in both accurateness and quadratic computing time.
Document Type: Thesis
File Description: application/postscript
Language: English
Relation: http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.16.4732
Availability: http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.16.4732
http://www.iti.upv.es/~prhlt/PAPERS/papers96-00/2000/Mollineda00e.ps.gz
Rights: Metadata may be used without restrictions as long as the oai identifier remains attached to it.
Accession Number: edsbas.11AF8864
Database: BASE
Description
Description not available.