Bayesian network is a robust structure for representing knowledge containing uncertainties in a knowledge-based system. In applications of expert systems and knowledge-based systems, it often happens that initial data are not sufficient to derive a conclusion of high enough certainty. Inference-guiding is in that case to identify the missing information, pursue its value, and lead inference to a conclusion. This paper presents and characterizes a criterion for effectively selecting key missing information, and thereby develops a “smart” inference approach with the inference-guiding function based on the newly developed criterion for uncertain inference in a Bayesian knowledge-based system.
"Inference-Guiding on Bayesian Knowledge-Based Systems,"
Journal of International Technology and Information Management: Vol. 18:
3, Article 9.
Available at: https://scholarworks.lib.csusb.edu/jitim/vol18/iss3/9