Breast cancer clustering using modified spherical K-Means

Abstract

Abstract Clustering is one of common techniques to group dataset into subsets based on distance measure. It has been applied in machine learning, pattern recognition, data mining, image analysis, and bioinformatics. Spherical k-means is one of clustering methods to address computational efficiency and solution quality in terms of deciding an action. In this paper, we used modified spherical k-means by using kernel radial basis function (RBF) by inner product measures in spherical k-means to cluster breast cancer Coimbra dataset from UCI machine learning into clusters. A new clusters will defined to healthy control cluster and patient cluster based on medical records. The highest accuracy results of kernel spherical k-means (SPKM) clustering method with radial basis function (RBF) kernel in breast cancer Coimbra (BCC) dataset is 72,41%. Addition of kernel to spherical k-means makes the results of accuracy be stable than using spherical k-means.