Enhanced AI Based Diabetic Risk Prediction Using Feature Scaled Ensemble Learning Technique Based on Cloud Computing

Diabetes causes many health problems including microvascular disease, macrovascular abnormalities and neuropathy. Diabetes is one of the costliest diseases from an economic perspective, and most adults through diabetes live in small- and intermediate-revenue countries, creating additional financial burdens and problems for these countries.The healthcare industry is one ofvast industries in the comprehensivebudget with revenue and investment increasing every year. Early and exactanalysis of diabetes, especially in the initial stages of the disease, can be difficult for medical professionals. To resolve the issue, we deploy the AI based ensemble learning method. To attain the objective, we perform three stagesfor instance preprocessing, feature selection, and classification. In the first phase, we preprocess the diabetes dataset by using numeric normalization method. This method is used for the diabetes dataset improves the validity of subsequent ML progression and produces more reliable and descriptive results. After preprocessed the dataset we go to select the features based on Correlation-based feature selection (CFS) method. CFS aims to select a feature subset with strong predictive power while minimizing redundancy by estimating the correlation between features and target variables and cross-correlation between features. The proposed method needs to select the important features for classification. At last, we classify the diabetes dataset on the basis ofreviewed ensemble learning methodologies to progress prediction accurateness. A comparison of the review methods for diabetes risk prediction includes K-Nearest Neighbor (K-NN), Whale Optimization and Fuzzy Neural Network (WO-FNN), Decision Tree (DT), and Support Vector Machine (SVM). In this paper, we use K-NN, WO-FNN and DT as the base models and SVM as the meta-model. In the ensemble learning methodology we examine that SVM is the best method for forecasting the diabetes at premature stages by the diabetes dataset compare with other methods. The SVM approach provides a robust and accurate method for diabetes risk prediction based on diabetes datasets. Its capability to operate complex, distinguished-dimension data, combined with its resilience to overfitting, makes it a valuable tool in the field of clinical data analysis.