Design of an ANFIS based Decision Support System for Diabetes Diagnosis

Designing of an efficient medical decision support system is crucial and plays an important role in early prediction of a disease and assists specialists with appropriate medication. Diabetes Mellitus is one of the chronic diseases which is characterized by increased levels of glucose in blood due to defects in insulin secretion in the body. This could lead to long- term damage and failure of different organs especially eyes, nerves, kidney and heart. Thus, an early detection and proper medication could reduce the risk of these complications. Researchers have proposed several methods for diagnosis, one of which is the fuzzy based expert system suitable for diagnosis in medical domain considering ambiguous clinical data. Adaptive Neuro Fuzzy Inference System (ANFIS) is proposed in the following research paper, in order to improve the classification accuracy by considering the adaptability factor through neural networks integrated with fuzzy system. The Pima Indian Diabetic Dataset (PIDD) was trained and tested for classification in MATLAB. The results show that ANFIS based classification gives better accuracy when compared with other methods.