Application of Fuzzy Kernel K-Medoids Method for Cancer Classification based on Metal Concentration in Blood

Abstract

Classification technique has already been applied widely in the medical data. One of its applications is for classification of cancer. The accuracy of this technique highly depends on the type of data to be processed (whether the data are separable or non-separable) and the dissimilarity function used. To surmount those hindrances and to improve the accuracy of classification therefore a method named Fuzzy Kernel K-Medoids (FKKM). The method can be used for separable or non separable of data. Based on the research on the concentration data of Zn, Ba, Mg, Ca, Cu, and Se in blood in order to diagnose cancer, FKKM gives better result than the Support Vector Machines Method. This paper will discuss an application of the FKKM method on the concentration data of Zn, Ba, Mg, Ca, Cu, and Se in blood samples and compared with the Support Vector Machines Method for the diagnosis of cancer. Results showed that the FKKM method produced a better result than the Support Vector Machines Method.