A practical criterion for the success of a knowledge-based problem-solving system is its usefulness as a tool to those working in its specialized domain of expertise. Here we describe several applications of the PROSPECTOR consultation system to mineral exploration tasks. One was a pilot study conducted for the National Uranium Resource Estimate program of the U.S. Department of Energy. This application estimated the favourability of several test regions for occurrence of sandstone uranium deposits. For credibility, the study was preceded by a performance evaluation of the relevant portion of PROSPECTOR's knowledge base, which showed that PROSPECTOR's conclusions agreed very closely with those of the model designer over a broad range of conditions and levels of detail.
A practical criterion for the success of a knowledge-based problem-solving system is its usefulness as a tool to those working in its specialized domain of expertise. This paper describes an evaluation and several applications of a knowledge-based system, the PROSPECTOR consultant for mineral exploration. PROSPECTOR is a rule-based judgmental reasoning system that evaluates the mineral potential of a site or region with respect to inference network models of specific classes of ore deposits. Knowledge about a particular type of ore deposit is encoded in a computational model representing observable geological features and the relative significance thereof.In Hayes, J. E., Michie, D., and Pao, Y.-H. (Eds.), Machine Intelligence 10. Ellis Horwood.