Abstract
The analysis of epidemiological data is critical to disease prevention and control programs geared toward improving, promoting, and protecting the health of communities. Various decision-making support systems have been modelled using artificial neural networks and fuzzy inferences. A neuro-fuzzy inference system based on the Takagi-Sugeno system was developed in the early 1990s that integrates the advantages of neural networks with fuzzy logic principles, such as self-learning and knowledge representation. Adaptive neuro-fuzzy inference systems are devised and evaluated here as means of characterizing the severity of a laboratory-confirmed COVID-19 case. The authors describe the underlying architecture for ANFIS with various clustering approaches, including grid partitioning, subtractive clustering, and fuzzy c-means. A total of 385 cases with eight potential predictors is used to develop, validate, and evaluate the model.