Fuzzy optimized classifier for the diagnosis of blood pressure using genetic algorithm

Abstract

We propose to optimize the fuzzy rules, which are based on an expert, the objective is to classify the blood pressure level in a correct way with the necessary number of rules and not to have some type of mistake at the moment of giving the diagnosis, since the use of unnecessary rules could cause a confusion in the fuzzy classifier. The fuzzy classifier is only part of the neuro fuzzy hybrid model, which uses techniques such as: neural networks, fuzzy logic and evolutionary computation, in this latter technique, genetic algorithms are used, which use individuals as possible solutions and thus obtain the best solution, in this case find the appropriate number of fuzzy rules for fuzzy system. This study aims to model blood pressure for 24 h and obtain the trend per patient, once this trend is obtained, this information enters a fuzzy system based on rules given by an expert, who will be classified into some of the blood pressure levels based on level European guide and finally give us a diagnosis per patient.