Search

The paper introduces the modular approach in the Characteristic Objects method for the reduction of the rules base and thereby avoiding the curse of dimensionality. In this way, the required number of comparisons, between characteristic objects, is highly reduced. The modular approach decomposes a decisional model to related submodels, where each submodel aggregates correlated criteria. In the result, the modular solution consists of several fuzzy modules, which create a hierarchical structure of the considered problem. The proposed approach is compared with the conventional approach by using two exemplary decision functions. For this purpose, three and six-dimensional reference functions are used. Performed experiments enabled the comparison of an accuracy a proposed approach. [ABSTRACT FROM AUTHOR]

This paper introduces the Characteristic Objects Method (COMET) as a useful multi-criteria decision-making (MCDM) method for use in medical applications. The author presents an assessment of the risk of hard arteriosclerotic cardiovascular disease (ASCVD) events. First, an MCDM mini model will be obtained using the COMET, this mini model is the easiest to identify but has the lowest accuracy. The mini model uses only two values for each criterion as characteristic values. For the ASCVD problem, the following four criteria will be used: age, total cholesterol, HDL cholesterol and systolic blood pressure. The author assumes that a white, male patient who is nonsmoker and is free of diabetes or hypertension. The model will next be used to build a ranking for example patients. Finally, the obtained results will be verified using the statistical data (ACC/AHA model) to check the accuracy of the COMET procedure. [ABSTRACT FROM AUTHOR]