Machine Learning Solutions for Healthcare Problems

Abstract

Machine learning (ML) is a multidisciplinary field concerned with answering how we can construct computer programs that automatically improve the experience. ML footprints can be found in almost any field, including healthcare. In this dissertation, specific analytical solutions are developed for healthcare problems using a combination of ML and Optimization (Opt) tools. This presentation includes two research questions in healthcare and our proposed solution to address them. In the first study, we explored how the changes in the amino acid sequences of the HIV-1 Env can be predicted using the neighboring variability. This study is a vital step toward predicting future changes in the amino acid sequences and personalized treatments for HIV patients. In the second study, we proposed an algorithm to accurately predict the ICU needs of COVID-19 patients so that the resources are managed more efficiently, and more lives could be saved.