Automatic Extraction of Metadata Based on Natural Language Processing for Research Documents in Institutional Repositories

Herrera-Urtiaga A.P.; Torres-Cruz F.; Mendoza-Mollocondo C.I.; Paredes-Quispe J.-R.; Chambi-Mamani E.W. · Smart Innovation, Systems and Technologies (2023) · DOI: 10.1007/978-3-031-31007-2_19

Resumen

The Institutional Repositories allow to organize, preserve, and disseminate scientific and academic production of an institution. Metadata are data about other data, which help identify, recover, and preserve the information of the research documents produced by the Universidad Nacional del Altiplano. The extraction of metadata is a time-consuming process in the treatment of research documents, and there may be a risk of introducing erroneous information. The purpose of this research is to optimize the extraction of metadata on 380 research documents, which is a necessary process for their preservation, Therefore, an algorithm was developed to allow the automatic extraction of metadata using natural language processing (NLP) techniques for Institutional Repositories (E-MeRI), to which the linear algorithmic complexity O(n) was determined, and at the same time, the level of accuracy was determined between 96% and 99% of correct results based on the metrics Precision and Recall. Finally (E-MeRI) proved to be efficient compared to other extractor tools with 50 test documents.

Instituciones: Universidad Nacional del Altiplano