Minnesota code: A neuro-fuzzy-based decision tuning

Abstract

The Minnesota Code is the evaluation method of reference ECG signals. The experimental studies compare the effectiveness of the computer based Minnesota Code applications to human usage of the code system, and the results showed that computers are as effective in the evaluation of ECG signal with the Minnesota Code as humans are with visual analysis. A fuzzy-based approach can be used to bypass known imperfections and imprecision of the existing Minnesota Code rules. A fuzzy-based approach also has issues with corner case inputs, which can lead to incorrect partial results and incorrect diagnostics outputs. The fuzzy environment provides more information for the medical expert or for the further levels of the whole hierarchically organized diagnostic structure. The authors of the paper present a possible solution for fine-tuning the diagnostic rules using neural networks. In this paper, the standard fuzzy-based approach is extended to a neuro-fuzzy solution.