Comparative study of adaptive neuro-fuzzy and fuzzy inference system for diagnosis of hypertension

Abstract

Adaptive neuro-fuzzy inference system (ANFIS) for diagnosis of hypertension is proposed in this paper. Hypertension patient's records are collected under the supervision of physician on real time basis. The system is trained using hybrid training algorithm. The trained model is tested by using new patient's record. Further the performance of the diagnosing capability of the system is compared with judgment of physician on same test cases. The performance of the system is also compared with existing fuzzy expert system in terms of several parameters. We have obtained 94.63% accuracy from the experiments made on record collected from physician and it was very promising with regard to existing system. We have obtained sensitivity 97.50%, specificity 93.33% and precision 98.11% values for diagnosis hypertension.