This paper describes a method for optimizing the cost matrix of any approximate string matching algorithm based on the Levenshtein distance. The method, which uses genetic algorithms, defines the problem formally as a discrimination between a set of classes. II is tested and evaluated using both synthetically generated strings of symbols and chain code data extracted from the international Unipen database of on-line handwritten scripts. Experimental results show that this approach can effectively discover the hidden costs of elementary operations in a set of string classes. (C) 1998 Pattern Recognition Society. Published by Elsevier Science Ltd. All rights reserved. [References: 14]
展开▼