Fuzzy-based procedures of multiparametric pattern recognition have been established in technical diagnostics and industrial automation for more than 10 years. The consideration of fuzziness has broken new ground in computer-aided decision support which still goes on successfully today. The capability of this methodology significantly benefits from the inclusion of knowledge, experience, and applied examples. A further advantage of these procedures consists in the terminological transparency and ease of comprehension. Instead of abstract mathematical formulas, the algorithms are described by verbal rules. This aspect particularly facilitates the interdisciplinary cooperation between users and mathematicians, and in this way facilitates a time- and cost-effective system development as well.
However, the utilization of fuzzy systems is effectual only if the problem is suitable for fuzzy treatment. As an example, these requirements are examined in the analysis of tumor marker profiles. The illustration of basic ideas and principles provides the understanding of the mathematical data evaluation without direct confrontation with algorithms. As a result, an increase of sensitivity on high specificity is achievable. The particular steps of calculation seem to conform to the subjective decision making when tumor markers are analyzed.
Print ISSN: 0025-8466
Volume: 28, 04/2004
Pages: 116 - 121