Type-2 Diabetes Mellitus Prediction Through Ensemble Learning Technique Based on Gene Data and Machine Learning Approach

Background: Diabetes is a metabolic disorder disease that affects 453 million people worldwide. Furthermore, many people living with this disease do not realize the serious effects on their health status. Inadequate knowledge and late diagnosis bring severe problems and so many deaths happen each year; the development of timely detection and diagnosis of diabetes is essential. Objective: The study aims to use machine learning approaches analysis of diabetes. That makes the diabetes management easier. The advancement for disease detection and understanding of symptoms are rapidly increasing in developing fields. Method: Disease prediction is one of the captious tasks. There are many cardiovascular diseases spread around the world like diabetes and cancer. The machine learning techniques have been successfully applied in assorted applications of medical diagnosis. Result: 99% accuracy was obtained from using bagging classification techniques on the gene dataset. Previously when SVM classification was applied to the gene dataset it provided an accuracy of 92.5%. Conclusion: The bagging classification technique is used for data preprocessing and classification to test and train the gene dataset in diabetes detection over performing the best from SVM, KNN, and random forest.