A Neuro fuzzy hybrid model (NFHM) is used as a new Computational Intelligence method for blood pressure (BP) classification. The objective is to model the behavior of blood pressure based on monitoring data of 6 days with 4 samples per day, two samples were taken at noon and two more in the afternoon from a set of patients. The new model uses the data as input to the classifier to estimate the level to which the patient belongs.Design and Method:
Intelligent computing techniques, such as neural networks and fuzzy logic, working with a modular architecture are used to model the BP behavior in patients. We consider different number of layers and neurons with different learning parameters and we also use a fuzzy rule base with expert knowledge of BP classification so we can achieve a more accurate modeling.Results:
Of the 30 monitored patients in the study; these were considered in different stages, such as high BP, normal BP and low BP, this is to test our system with patients in different conditions. Different architectures of NFHM were used to achieve better results: Architecture 1 produced better results than Architecture 2 in all 30 experiments. First we have the systolic architecture with a Mean Error (ME) of 3.747% and we have the diastolic architecture with a ME of 2 %.Conclusions:
This type of NFHM actually simulates the human Medical reasoning by using a set of decision rules, in this case of BP classification. This is a very efficient, less time consuming and more accurate method to classify BP. Finally we can note that is an effective diagnosis method for hypertension or hypotension, which can help a physician to achieve better accuracy when providing a diagnosis to the patient.