LexiCube: Multidimensional Generation and Analysis of News with LLMs and OLAP

Autores/as

  • Rodrigo Nogueira Federal University of Santa Catarina
  • Ian Santos Instituto Federal de Educação, Ciência e Tecnologia Catarinense
  • Rafael De Moura Speroni Instituto Federal Catarinense
  • Daniel Fernando Anderle Instituto Federal Catarinense
  • Douglas Hörner Instituto Federal de Educação, Ciência e Tecnologia Catarinense

Palabras clave:

LLMs, OLAP, Data Warehouse, News, Knowledge Extraction

Resumen

The continuous growth of digital news production presents challenges for organizing and analyzing large volumes of data. This paper introduces LexiCube, an automated system for collecting, transforming, and structuring journalistic content into a multidimensional analytical model. The approach integrates ETL techniques, journalistic ontologies, and large language models (LLMs), aiming to create a repository suitable for in-depth analysis and interpretation.

The methodology includes automated text extraction from news portals, processing in a staging area, and loading into a textual data warehouse. The data is semantically enriched and converted into vector representations, enabling automatic categorization, temporal analysis, and named entity extraction.

Among the results obtained are the generation of knowledge graphs, identification of lexical patterns over time, and thematic organization of the content. These elements support more contextualized analyses grounded in textual data.

The findings indicate that LexiCube contributes to the transformation of journalistic data into structured knowledge, allowing for the visualization, segmentation, and interpretation of information from multiple perspectives. The system offers potential applications in environments that require continuous monitoring and analysis of unstructured content.

Descargas

Los datos de descargas todavía no están disponibles.

Publicado

2025-12-05

Cómo citar

Nogueira, R. ., Santos, I. ., De Moura Speroni, R. ., Anderle, D. F. ., & Hörner, . D. . (2025). LexiCube: Multidimensional Generation and Analysis of News with LLMs and OLAP. Congreso Internacional De Conocimiento E Innovación - Ciki, 1(1). Recuperado a partir de https://proceeding.ciki.ufsc.br/index.php/ciki/article/view/1733