Case-based reasoning shell for complex databases

Licensing opportunity

Researchers in the Artificial Intelligence Applications Institute at Edinburgh University have developed a case-based reasoning shell that can be applied to large and complex databases.

The researchers say the technique of case-based reasoning is particularly applicable to problems where the domain is not understood well enough for a robust statistical model or system of equations to be formulated. While the general case-based reasoning methodology can be applied across different application domains, the implementation of the case retrieval and similarity scoring functions is typically highly-customised to the problem at hand.

Case-based reasoning rests on the insight that new problems are often similar to ones encountered previously, and, therefore, that past solutions may be of use in the current situation. This means that commercial and industrial applications can be developed rapidly using existing corporate databases as knowledge sources.

The Edinburgh case-based reasoning shell has an advanced performance optimisation module, based on a genetic algorithm and is implemented in Java.

Case-based reasoning has the advantage that it is an efficient, robust and readily-understood technique for data analysis. It can be used to convert existing databases into exploitable knowledge, assessing the underlying patterns and uncovering knowledge implicit in datasets, which is critical to improving products.

Possible applications areas include help desks, where it can be used to search for past solutions from corporate databases; in the pharmaceutical industry, where cased-based reasoning could be applied to diagnosis/treatment decisions, to deciding whether or not a particular treatment is appropriate, given past cases, in analysing data from clinical trials; for protein structure prediction; or in fraud detection to reduce the number of applications referred for expert investigation.

Edinburgh University is looking for industrial partners interested in licensing the core case-based reasoning engine technology.

For more information, see the project page at: http://www.university-technology.com/details/case-based-reasoning---learning-through-experience