The authors identify and discuss a generic induction approach called interactive induction, which offers an alternative method for eliciting subjective information from an expert during the course of knowledge acquisition. The approach extends pure induction by involving the expert in the provision of additional subjective knowledge and in the incremental evaluation and validation of knowledge induced. It is still able, in principle, to induce knowledge beyond that articulated or known by the expert. Moreover, unlike earlier induction methods, interaction does play a key role; induction is not viewed as an automatic process. Consideration is given to the various aspects of the knowledge engineering problem that are relevant to the interactive induction approach and software features that are needed to support it. Relevant theory is discussed along with lines for future research.<<ETX>>
[1]
Wray L. Buntine.
Decision tree induction systems: A Bayesian analysis
,
1987,
Int. J. Approx. Reason..
[2]
Max Henrion,et al.
An Experimental Comparison of Knowledge Engineering for Expert Systems and for Decision Analysis
,
1987,
AAAI.
[3]
Stephen Muggleton,et al.
Structuring Knowledge by Asking Questions
,
1987,
EWSL.
[4]
Chris Carter,et al.
Assessing Credit Card Applications Using Machine Learning
,
1987,
IEEE Expert.
[5]
Ross D. Shachter.
DAVID: influence diagram processing system for the macintosh
,
1986,
UAI.