Applying Bibliometrics and a RoBERTa Transformer in the Circular Bioeconomy: A PRISMA 2020 Systematic Review

This exploratory methodological study demonstrates an integrated workflow that combines systematic evidence collection Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA 2020), bibliometric mapping, and Transformer-based natural language processing (RoBERTa) to generate multi...

Descripción completa

Guardado en:
Autores principales: Gary Christiam Farfán Chilicaus, Alexander Fernando Haro-Sarango, Angela Fremiot Rodriguez-Armas, César Augusto Herrera-Asmat, Silvia Cachay-Salcedo, Rosa Amable Salcedo-Davalos, Violeta Claros Aguilar De Larrea, Emma Verónica Ramos Farroñán
Formato: Artículo de revisión
Lenguaje:inglés
Publicado: 2026
Materias:
circular bioeconomy
circular economy
science mapping
RoBERTa
natural language processing
sentiment analysis
hedging
discursive framing
research trends
conceptual framework
measurement gaps
epistemic bias
scientometrics and bibliometrics research
Bioeconomy and Sustainability Development
Meta-analysis and systematic reviews
Acceso en línea:Enlace permanente (DOI o Handle)
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
Sumario:This exploratory methodological study demonstrates an integrated workflow that combines systematic evidence collection Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA 2020), bibliometric mapping, and Transformer-based natural language processing (RoBERTa) to generate multi-layer insights from Circular Economy-related scholarship, using circular bioeconomy literature as a domain case (2017–2025). Searches across Scopus, ScienceDirect, Taylor & Francis, and SAGE retrieved 2643 records; after deduplication and screening, 50 studies were included (mean quality 13.2/16; 68% high quality). Bibliometric mapping (VOSviewer; Scopus subset n = 1468) revealed three thematic clusters that converge with five conceptual framings extracted via qualitative synthesis, providing cross-method validation of the pipeline’s structural and interpretive outputs. The NLP layer identified a predominantly positive discursive valence in the English-language title–abstract–keyword corpus derived from Scopus records, with declining polarity and increasing subjectivity over time. Because these estimates were obtained from composite bibliographic text fields rather than full-text discussion sections, they should be interpreted as indicators of narrative framing rather than as direct evidence of epistemic bias or empirical overstatement. Within that scope, the joint reading of polarity, subjectivity, hedging, and measurement gaps suggests a possible mismatch between acknowledged evaluative limitations and the caution used to communicate them.

Ejemplares similares