Feature selection from large-scale radiomics data: Application to heart disease diagnosis

Abstract

Radiomics have become in the past years one of the most interesting fields to be studied and analyzed in medicine. Since their first implementations, researchers have been trying to design new algorithms to develop this tool in order to predict with higher accuracy the appearance of some diseases like cancer. In this thesis we will focus in a method for feature selection, one of the most challenging questions that come up when working with radiomics. The aim of this new algorithm is to improve the accuracy and efficiency by making use of parallel computing. We will face, as well, one of the main issues of the new century, dealing with the Big Data.