Support vector machine classification learning algorithm for diabetes prediction

Abstract

Diabetes is actually one of the primary causes of human mortality. Diabetes is an intense disease affecting various parts of the human body. Diabetes can rise long-range complications including, renal failure and cardiac failure. It is therefore imperative that diabetes be diagnosed in a timely manner to people all over the world. This study develops a method for diabetic classification using machine learning techniques. In this study, Support Vector Machine (SVM) is employed to classify the diabetic disease into two classes based on its different functions, namely, linear, polynomial, and sigmoid functions. The evaluation performance of this study is performed before and after applying the pre-processing stage using different standard criteria. The higher results were obtained by polynomial function 83.77% for accuracy, 86.07% for sensitivity, and 81.97% for specificity. Finally, a comparison between this study and some of the previous studies was addressed, based on the comparison it is shown that this study has a better ability to classify diabetic disease than previous studies.

Keywords

• classification
• data analysis
• diabetic
• feature extraction
• machine learning
• pima dataset
• pre-processing
• svm technique