Error Detection, Recovery, and Repair in the Translation of Inorganic Nomenclatures. 2. A Proposed Strategy
Irene Luque Ruiz, J. L. Cruz Soto, Miguel Ángel Gómez‐Nieto · Journal of Chemical Information and Computer Sciences · 1996
The repair of lexicographic errors occuring during the computer translation of knowledge is performed by approximate string matching algorithms. This article presents a new algorithm that attempts to repair lexical errors in inorganic chemical names. The algorithm allows the repair of an incorrect string containing one or more errors due to deletion, insertion, or transposition of multiple characters. Repair may be carried out on all or only part of the incorrect string, so the algorithm may be used to repair strings containing more than one incorrect morpheme. A model is proposed for the use of this algorithm in the repair of lexical errors in inorganic chemical names. The model attempts to generate a set of possible repair strings free from either syntactic or lexicographic errors. Repair is performed by comparing all or part of the error string with the set of terminal symbols or morphemes included in the grammar. A new model is proposed for the calculation of similarity; it is highly discriminating and allows the repair of parts of an error string.