KLASIFIKASI PASIEN DIABETES MELLITUS MENGGUNAKAN METODE SMOOTH SUPPORT VECTOR MACHINE (SSVM)

Rizky Adhi Nugroho, Tarno Tarno, Alan Prahutama

Abstract


Diabetes Mellitus (DM) is a high-risk metabolic diseases. Laboratory tests are needed to determine if the patients suffer from a Diabetes Mellitus. Therefore, it needs a classification methods that can precisely classify data according to the classes criteria. SVM is one of commonly used methods of classification. The basic concept of SVM is to find the bes separator function (hyperplane) that separates the data according its class. SVM uses a kernel trick for nonlinear problems, which transforms data into high-dimensional space using kernel functions, so it can be classified linearly. This research will use a developed methods of SVM called SSVM, that adds smoothing function using Newton-Armijo method. The smoothing methods is needed to correct the effectiveness of SVM in big data classifying. The result is indicating tha SSVM classification prediction using Gaussian RBF kernel function, can classify 98 out of 110 patient data of Diabetes Mellitus correctly according the original class.

Keywords : Diabetes Mellitus, Classification, Support Vector Machine (SVM), Smooth Support Vector Machine (SSVM), Kernel Gaussian RBF.


Full Text:

PDF

Refbacks

  • There are currently no refbacks.



Creative Commons License
Jurnal Gaussian by Departemen Statistika Undip is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.

Flag Counter