In terms of knowledge reusability, analogical reasoning gives more flexibility when applying knowledge to new situations. In general terms analogy builds a mapping between old cases and a new one, in other words mapping the features of the target new case to the source.
The basis of analogy is that if two situations (even objects) have a thing or a feature in common then they are likely to have other things in common. For example an aeroplane and a ship both carry passengers, other things they have in common are that they both travel from one place another, they both carry cargo, they both have a captain and a crew.
Hall(1989), Kedar-Cabelli(1988) and Wolstencroft(1989) have all proposed a similar framework for analogical reasoning. The following represents a typical framework for analogical reasoning:
|
The problem of analogical retrieval involves selecting the potential features of the target and source that increase the retrieval of useful source analogy and indexing knowledge on the basis of those features. |
|
Upon retrieving the relevant source, it is often necessary to determine additional features and relations of source. In some cases developing a specific problem solving strategy (or explanation) in the source domain for the target domain might be inevitable. |
|
This stage involves building a map from the features of source to target domain. |
|
This stage examines if the mapping between source and the target is valid, sometimes it requires some modification of the mapping. |
|
This is the stage of storing the gained knowledge for future use. |
There are two analogical problem solving strategies these are:
Authored by Serengul Smith
E-mail to:
serengul1@mdx.ac.uk
School of Computing Science Middlesex University
Revised: September 1998