Abstract
Having got a working skeletal expert system and some rules for our application, we are faced with the inevitable question: does it work? This, again, is a question that expert systems pose in common with conventional computer systems. In just the same way that conventional applications need to be validated, so does ours. In fact, the need for validation is even greater. Conventional applications start on much firmer ground; the algorithms that they implement rest on scientific laws, established professional or business practices, or equally well-assured grounds. In contrast, what we are implementing is far less assured, resting mainly on heuristics, rules of thumb or, at most, formulations borne out by practice but never subjected to rigorous examination before.
Preview
Unable to display preview. Download preview PDF.
Further reading
The nature, process, and verifiability of judgements is laid bare in the book by Sir Geoffrey Vickers, The Art of Judgement: A Study in Policy Making (University Paperbacks, Methuen, 1968) and should be made compulsory reading for prospective knowledge engineers.
Copyright information
© 1985 Peter S. Sell
About this chapter
Cite this chapter
Sell, P.S. (1985). Validation. In: Expert Systems — A Practical Introduction. Macmillan Computer Science Series. Palgrave, London. https://doi.org/10.1007/978-1-349-07416-7_6
Download citation
DOI: https://doi.org/10.1007/978-1-349-07416-7_6
Publisher Name: Palgrave, London
Print ISBN: 978-1-349-07418-1
Online ISBN: 978-1-349-07416-7
eBook Packages: Computer ScienceComputer Science (R0)