Information Technology Reference
In-Depth Information
successfully to domains struck by the “curse of dimensionality” [45]. This
“curse,” a well-known issue in bioinformatics, refers to the availability of a rela-
tively small number of cases, each having thousands of features. In their Gene-
CBR system, for cancer diagnosis, a case has 22,283 features, corresponding to
genes. The authors have designed a hybrid architecture for Gene-CBR, which
combines fuzzy case representation, a neural network to cluster the cases for ge-
netically similar patients, and a set of if-then rules extracted from the case base
to explain the classification results [45]. In other work, Montani explains how
CBR can be used to configure the parameters upon which other AI methodolo-
gies rely [79]. McSherry presents a novel hypothetico-deductive CBR approach
to minimize the number of tests required to confirm a diagnostic hypothesis [80].
Still other work capitalizes on the complementarity among knowledge bases, on-
tologies, and CBR [81].
8 Summary and Conclusion
Biomedical domains have long been a focus of AI research and development.
CBR is an AI approach that is especially apropos, due to its leverage of ex-
periential knowledge in the form of cases. Health sciences professionals reason
naturally from cases, which they use, among other things, to inform practice,
develop clinical guidelines, and educate future professionals. CBR is also highly
synergistic with other AI approaches, which may be combined with CBR in mul-
timodal reasoning systems. CBR in the Health Sciences systems have been built
to support diagnosis, classification, treatment planning, and education, in areas
as diverse as oncology, diabetology, psychiatry, and stress medicine. They sup-
port biomedical research and bioinformatics, as well as clinical practice. CBR in
the Health Sciences researchers continue to branch out into new reasoning tasks
and additional biomedical domains.
CBR and the health sciences impact each other in mutually beneficial and
synergistic ways. The health sciences provide a rich and complex experimental
field for evaluating CBR methodologies and extending the CBR approach. A
complementarity with statistics allows CBR to support evidence gathering for
biomedical research and evidence-based medical practice. As computer science
becomes more application driven, and as the health sector continues to expand,
this interdisciplinary field is expected to grow. CBR in the Health Sciences is
poised today to push the state of the art in AI while contributing clinically useful
systems that promote health and wellbeing.
References
1. Holt, A., Bichindaritz, I., Schmidt, R., Perner, P.: Medical applications in case-
based reasoning. The Knowledge Engineering Review 20(3), 289-292 (2005)
2. Bichindaritz, I., Montani, S., Portinale, L. (eds.): Special issue on case-based rea-
soning in the health sciences. Applied Intelligence 28(3), 207-285 (2008)
3. Bichindaritz, I. (ed.): Special issue on case-based reasoning in the health sciences.
Artificial Intelligence in Medicine 36(2), 121-192 (2006)
Search WWH ::
Custom Search