Adaptive algorithm in glossary search

Abstract

This paper deals with algorithm to curtail the computation time of approximate string matching by dynamic programming. We formulate the user requirement for eligible matches as inequality solving problem. The pith can be divided into two parts. First, the record is examined if has potential to fulfill the requirement before infiltrating into a time-consuming search. Then, only the region which can lead to the calculated similarity score is explored. This strategy shows a several fold speedup in a common glossary database.