This thesis explores the application of the Literature-Based Discovery (LBD) methodology in combination with Large Language Models (LLMs) to identify new solutions in the public sector. The main objective was to examine whether generative artificial intelligence can support the recognition of connections between problems, their characteristics, and potential interventions. For this purpose, I developed the AEI-NSPS application, based on the LBD ABC model, which enables both open and closed discovery, with results presented through interactive graphs. Empirical case studies from medicine and public administration, such as migraine and graduate unemployment, demonstrated that LLMs successfully generate meaningful hypotheses and connect them to existing literature with the support of external databases such as Scopus and PubMed. The findings indicate that the LBD methodology can be transferred beyond medicine and applied as a tool to support innovation in the public sector.
|
