Diabetes Diagnosis with Intelligent Optimization Based Support Vector Machines

Artificial Intelligence is one of the most important scientific fields that can be applied effectively in different real world problems and has been shaping our future for a long time. While there are various types of problems in which it is applied, medical diagnosis is one of the most remarkable one among them. Moving from the explanations, objective of this study is to realize diabetes diagnosis by using intelligent optimization based Support Vector Machines (SVM). In this context, five ones of today's recent intelligent optimization algorithms were used for optimizing a non-linear SVM, which is using a Gaussian-RBF kernel function. Obtained findings showed that hybrid systems formed with different algorithms show different-level success but in general, good level accurate results can be achieved via intelligent optimization-SVM approach. Hereby, the study also confirms that this followed approach has a significant potential for Artificial Intelligence based diagnosis.