Desarrollo de un sistema Web basado en matrices de distancias para seleccionar artículos académicos similares en la revisión de la literatura científica

Abstract

Due to the constant increase in publications of scientific articles in various virtual libraries worldwide every year, searching for relevant academic documents for research has become a complex process that sometimes requires significant time and effort. This research article addresses this problem by developing a web system that makes it easier for users to identify the most relevant scientific articles for their research. The system allows users to upload a set of data that includes the attributes, titles, keywords, and abstracts of each extracted document. Each scientific article was subjected to the NLP process to purify and homogenize the information. To measure the divergence between scientific documents, similarity matrices are calculated using two different metrics that assemble the three attributes into a single weighted matrix. The results are presented through a web interface, including a summary table, a heat map, a twodimensional scatter diagram, and a graph. The experimentation was carried out using a data set composed of 192 scientific articles from Springer, IEEE, and Scopus, covering the areas of Exact Sciences, Computer Sciences, Medicine, and Social Sciences. The results revealed greater similarity between documents that share the same study topic.