PROSPECTOR was an expert system developed at Stanford in the
late 1970s to evaluate geological sites for potential mineral deposits. In 1984 it was
instrumental in discovering a molybdenum deposit worth 100 million dollars.
PROSPECTOR was based on a semantic network and used a combination of mostly forward chaining and some backward chaining. It adopted a form of depth first search guided by the probabilities of different hypotheses and used Bayes Theorem to deal with uncertainty. The dialogue component and explanation component used mixed control.
Defined on page 429