Compartir
Título
Revisión exploratoria de literatura científica en acuicultura: Análisis de tendencias utilizando un mocelo probabilístico bayesiano y herramientas de machine learning
Autor(es)
Director(es)
Materia
Minería de texto
Asignación latente de Dirichlet
Artículos publicados
Modelado de tópicos
Text mining
Latent Dirichlet Allocation
Published articles
Topic model
Clasificación UNESCO
1209 Estadística
Fecha de publicación
2020-07-20
Resumen
{EN}Research in aquaculture develops fast as it has to respond to the significant growth of this industry; there are many biological challenges, technical improvements and technology developments that need to be addressed by researchers. Thousands of articles on aquaculture have been published, so it is laborious and time consuming to extract information from accumulated collections. The aim of this study was to understand the distribution patterns and trends of the literature available in the field of aquaculture in order to improve knowledge, nature and structure of these publications. This study performed a literature review of 38319 abstracts published in 14 top-tier aquaculture journals, between the years 1972 and 2019. A Latent Dirichlet Allocation (LDA) was applied to perform text mining on the dataset, finding 40 key topics. Machine learning tools were used in the subsequent distribution and composition of words. As result, we found that topic modeling has the ability to segregate a collection of articles on different topics, and could be used as a tool to understand literature, not only recapturing known facts but also discovering other relevant topics. In general, the topics found confirm key areas of aquaculture research that have been identified by qualitative studies. However in our case it also provides a quantitative evaluation and analysis in the most recent scientific literature
URI
Colecciones