ABSTRACT
In this chapter, the authors present a conceptual framework for medical screening, with a case study focused on breast cancer diagnosis. Identification of those features that are of greatest interest is very much disease specific. The person might be viewed as a set of data and, by applying different methods, we open windows into this data. The reason for refining the data is that the information may be in different formats such as text, image, or sound, and must be transferred to some numerical representation before further processing in the decision step. In order to find a fully automatic π-function, we must extract the disease-specific information used by domain experts in their judgment. An interviewing approach, expressing the knowledge in linguistic variables is described in. The knowledge used by a radiologist when judging a mammogram must be described in the same way as we described the causes of breast cancer.
