Ensemble Learning Model for Diabetes Prediction

Diabetes, commonly referred to a group of metabolic diseases caused by chronically increased blood sugar levels are referred to as chronic disorders. Due to outliers (or missing values) and a minimal amount of labeled m data It is extremely difficult to make a robust and reliable diabetes prediction using diabetes datasets. A robust framework for diabetes prediction is proposed in this study by including feature selection, data normalization, outlier rejection, filling in missing values, and several ensembling strategies. The performance of the proposed framework for diabetes prediction outperforms the existing frameworks. Furthermore, it can produce superior results on the same dataset, which can enhance the accuracy of diabetes prediction. The public can access the source code for diabetes prediction.