A method for semi-automatic conceptual models representation from business documents written in natural language in Spanish

Abstract

Currently, the software development industry presents challenges for processing business information, especially that contained in textual documents. In the process of software requirements elicitation, business documents facilitate domain understanding and product evolution, although they are underutilized due to the associated costs and time. This paper proposes a semi-automatic method to extract information from business documents written in Spanish, using Natural Language Processing (NLP) techniques and representing it in conceptual models through class diagrams. The method was validated through a case study with software analysts, showing improvements in processing time while maintaining a similar level of interpretation compared to manual processes.