Decision Support System in Orthopedics Using Methodology Based on a Combination of Machine Learning Methods

Abstract

Analysis of data based on gene expressions characterizing serious disease is an area currently receiving high attention. The basic task is to classify patients, usually by searching for a small group of genes that provides sufficient classification power. However, very often, different gene combinations can describe different aspects of the problem being analyzed. In this paper, we present in a concrete example with one real dataset, a methodology that has repeatedly been successfully applied to different types of data. In addition to common statistical methods, this methodology combines methods such as a visualization of a dataset structure using networks, and feature-selection and neural network classification. The output of the application of the methodology is a system for decision support during the reoperation of patients with joint endoprosthesis.