Support vectors machine method andits software implementation for detecting cancer diseases

Abstract

The application of the method of support vectors for the classification of cells on tomographic images of human organs with signs of oncological diseases is considered. The equation of the dividing hyperplane and the width of the border between two reference vectors, which allow classification of cells into three groups corresponding to different types of diagnosis, including in the early stages of the disease, are given. The Soft Margin SVM was chosen for the software implementation of the considered method.