IFIBIO HOUSSAY   25014
INSTITUTO DE FISIOLOGIA Y BIOFISICA BERNARDO HOUSSAY
Unidad Ejecutora - UE
congresos y reuniones científicas
Título:
Text mining on indexed publications on life sciences signed by argentine authors
Autor/es:
JUAN JOSÉ CASAL; ROXANA TORIANO; RICARDO A. DORR
Reunión:
Congreso; REUNIÓN DE SOCIEDADES DE BIOCIENCIAS 2020; 2020
Resumen:
Europe PMC (ePMC) is an open science platform that allows access to worldwide Life Sciences publications, from trusted sources. The object of our study was the text of all scientific publications indexed in ePMC, signed by Argentine authors and published between 1955 and 2019 (inclusive). The objectives were: i) to statistically analyze the results of the search; ii) to discover, through data mining, non-explicit - often hidden - information structures and patterns in the text. The analysis was done through text mining (TM), developing automated workflows in the Knime software platform, and with semantic analysis tools as AntConc software. 75,294 articles were analyzed,published in 5,063 media, signed by 186,410 authors with workplace in Argentina or in collaboration with the country. We also worked with the text of 70,798 abstracts. TM allowed to extract information about journals, authors, and countries participating in the research, and the underlying information contained in titles and abstracts. The number of publications over time was correlated with Argentine economic parameters. The main publication media were detected, the number of authors signing each article was studied, and the countries sharing authorships with Argentina were analyzed. The pathologies that were mentioned in abstracts were detected; also the substances used for their treatment, grouping them by action site or by action mechanism. The topics that were especially covered by authors were found with unsupervised digital detection algorithms.The automated workflows (specially developed for this study) can be applied to other scientific or biomedical databases, to improve planification on research in prevention and treatment of illness. The unsupervised topic detection could serve to analyze decisions of authors on research subjects and to detect advances and deficits of an organized scientific policy.