Early Prediction of Diabetes and its Risk Factors based on ARIMA-ELMAN ANN Network

The prevalence of chronic diabetes is high. If diabetes could be predicted sooner, treatment results could be better.One such use of data mining techniques is in the field of disease prediction and early detection. To illustrate the potential value of these traits in predicting the presence or absence of diabetes, it is necessary to first explain the relationship between them. Technologies execute a wide variety of activities connected to diabetes, including clustering, association rule mining, and significant attribute selection determination. Priorities should be data preprocessing, feature selection, and model training. When it comes to data preprocessing, the cleansing and cleansing stage is in charge of dealing with errors, missing values, and inconsistencies. It takes advantage of the efficiency of the clustering algorithm and the reversing operation to exclude features with low F-scores. Prior to training ARIMA-ELMAN-ANN models, feature selection is necessary. The proposed approach clearly outperforms the two front-runners, ANN and ELMAN. Accuracy increased by 96.31 percent when the strategy was put into action.