Resumo

ALBO HERNANDEZ, Ramón Orlando et al. Requirements to improve the normalization of data in software for metric analysis of information. Rev. cuba. inf. cienc. salud [online]. 2018, vol.29, n.1, pp. 55-73. ISSN 2307-2113.

Due to the diversity of methods used to enter author-affiliation information, the resulting lack of standardization of bibliographic data has become one of the problems limiting analysis of metric information in terms of execution time, reliability of indicators and size of the data corpus. The purpose of the study was to propose requirements to improve data normalization in metric analysis software. To achieve this objective, a diagnosis was made of the main methods and techniques used worldwide in this type of study. The main result is the presentation of requirements to be met by an application for automated pre-processing of data for metric purposes. A proposal is made of the database, tasks, steps and algorithms that this application will contain. A combination of algorithms should be used to disambiguate author and affiliation fields.

Palavras-chave : data processing; data mining; bibliometrics; literature based discovery; data analysis.

        · resumo em Espanhol     · texto em Espanhol     · Espanhol ( pdf )