Development of Fuzzy Based Models for Detection System with More Accuracy for the Implementation of Healthcare/Clinical System

Many electronic infection detection systems often use binary classification techniques, which identify patient data as either pathology-indicating or normal in relation to different forms of infections. The Moni-ICU system, which is used in the critical care units of Vienna General Hospital, is noteworthy for introducing a novel third category called borderline. This categorization indicates patient data that partially satisfies the corresponding needs by closely aligning with virus-related clinical concepts or infections related to healthcare (HAI) surveillance categories. The probability of classifying patient data into normal, borderline, of pathological categories with respect to infection-related medical ideas and HAI definitions was calculated by integrating fuzzy sets of propositional fuzzy rules. This deviation from dichotomous approaches prevents the possible misunderstanding of borderline findings with normal ones, a major difficulty in binary classifications. Data collection included a thorough analysis of 5099 individuals treated over a two-year period, including adult patients across 10 ICUs at Vienna General Hospital. The Moni-ICU knowledge base, which includes fuzzy sets and rules for four top-level HAI definitions and 10 infection-related clinical concepts, made the categorization process easier. In order to verify the accuracy and dependability of the system, the assessment focused on how suitable borderline classifications were in relation to normal values for related HAI criteria.