versión impresa ISSN 0187-358X
Two algorithms to detect and solve normalization problems of author names in data originated in Thomson's ISI Science Citation Index are presented. The first algorithm allows detection of different names which could belong to the same person. The second one, based on the degree of similarity between two variants of the same name on a document, helps to determine whether two similar names correspond or not to the same person. In order to determine the efficacy of the algorithms, a control of normalized author data from a previous study has been used. The First algorithm detects 67% of name variants existing in the population under study, and the second one was successful in 74% of the cases.
Palabras llave : Author name normalization; Science Citation Index; Thomson ISI; Algorithms; Name variations.