J. Shavlik (1991).
Finding Genes by Case-Based Reasoning in the Presence of Noisy Case Boundaries. Proceedings of the DARPA Cased-Based Reasoning Workshop, pp. 327-338.

This publication is available in PDF and available in postscript.

Abstract:

Effectively using previous cases requires that a reasoner first match, in some fashion, the current problem against a large library of stored cases. One largely unaddressed task in case-based reasoning is the process of parsing continuous input into discrete cases. If this parsing is not done accurately, the relevant previous cases may not be found and the advantages of case-based problem solving will be lost. Parsing the data into cases is further complicated when the input data is noisy. This paper presents an approach to applying the case-based paradigm in the presence of noisy case boundaries. The approach has been fully implemented and applied in the domain of molecular biology; specifically, a successful case-based approach to gene finding is described. An empirical study demonstrates that the method is robust even with high error rates. This system is being used in conjunction with a Human Genome project in the Wisconsin Department of Genetics that is sequencing the DNA of the bacterium E. coli.

Return to the publications of the Univ. of Wisconsin Machine Learning Research Group.