A proposed approach for diabetes diagnosis using neuro-fuzzy technique

Diabetes is a chronic disease characterized by decrease in pancreatic insulin production. The immune system will be harmed due to this condition, which raise blood sugar levels. However, early detection of diabetes enables patients begin treatment on time, therefore reducing or eliminating the risk severe consequences. One most significant challenges healthcare unit diagnosis. Traditional techniques diagnosis are manual and prone inaccuracy. This paper proposed an approach for diagnosing using adaptive neuro-fuzzy inference (ANFIS) based Pima Indians dataset (PIDD). three stages pre-processing classification evaluation. Normalization, imputation, anomaly part stage. was done normalizing data, replacing missing values, local outlier factor (LOF) technique. In stage, ANFIS classifiers were trained hybrid learning algorithm neural network. Finally, evaluation procedures use last stage’s sensitivity, specificity, accuracy metrics. obtained 92.77%, it seemed rather promising compared other applications topic found literature.