Identification of Characteristic Descriptors and Creation of a Model for Predicting the Stage of Cancer

The paper presents an analysis of data for patients with registered oncological diseases and application of correlation analysis for detection of the most significant characteristic descriptors that have the greatest influence on determining the stage of malignancy. Based on the conducted theoretical studies herein are presented four groups of regression models created with three methods of machine learning - Decision Tree, Support Vector Machines and Ensemble algorithms - for three independent sets of data. A generalized model is also presented.