Human AI Construction of Bayesian Networks for Operational Decision Support -- A Virtual Survey Approach

arXiv cs.AI · 6 hours ago

Researchers proposed a methodology using large language models to construct Bayesian Belief Networks by having AI agents estimate probabilities across different personas, then applied trimmed-mean aggregation to reduce noise. The framework was tested on a model of customer intention to consult doctors in alternative healthcare, with 6 steps including parameter estimation and causal analysis. The approach bridges expert judgment and data-driven learning, revealing that subjective norms have stronger causal impact than self-efficacy on healthcare consultation decisions.