Linked Rule-Premises to Optimize the Fuzzy-Based Risk Assessment

In patient monitoring fuzzy-based evaluation can be used advantageously, because of its characteristics. This approach can work with input factors, for which sharp limits cannot be defined and it can handle uncertainties, subjectivity in the data as well as the evaluation process. The conventional Mamdani inference is the most suitable method, because it helps to build a model, which is much closer to human thinking, but its computational needs are very high, due to the complex-shape functions. This paper introduces a flexible patient monitoring system optimization, where the input functions can be tuned according to the personal statistics. Modified versions of the Mamdani inference are applied in this system, which can reduce computational complexity, while they retain the advantages of the original method.