This paper presents an approach for developing a Web-based decision support system that aids general practitioners in diagnosing dermatological problems using case-based reasoning. Our research aims to address the lack of decision support in current store-and-forward telemedicine systems and the need to deal with different costs of decisions. We propose a cost calculation method for minimizing the total misclassification cost. Two different techniques of measuring cost are evaluated. A series of tests is carried out on different sets of cost matrices for comparing the two methods. Our results show taking max values gives a lower misclassification cost and still manages to achieve a slightly higher accuracy compared to the summing method of instance-weighting.
CITATION STYLE
Ou, M. H., West, G. A. W., & Lazarescu, M. (2003). Dealing with decision costs in CBR in medical applications. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 2903, pp. 666–676). Springer Verlag. https://doi.org/10.1007/978-3-540-24581-0_57
Mendeley helps you to discover research relevant for your work.