Machine Learning Modelling

Machine Learning Modelling

 Nodira Nazyrova, Computing Department, Westminster International University in Tashkent

“A New Machine Learning Modelling Approach for Patients' Mortality Prediction in Hospital Intensive Care Unit”

Abstract: Machine Learning (ML) is a powerful tool in predictive modelling but subject to the problem of class imbalance. In this study, we tackle class imbalance with combining new features, data re-sampling, ensemble learning and an appropriate selection of evaluation metrics in a clinical setting. We built and evaluated 126 ML models to predict mortality in 48546 ICU admissions extracted from the Medical Information Mart for Intensive Care III (MIMIC-III) repository. Our approach has a considerable impact on the classification; it resulted in an improvement in the mortality status prediction. For evaluation, we implement a comparative multi-stage evaluation filter for a binary classification to compare all models. The best models are identified. The Area Under Receiver Operator Characteristic curves of the tested models range from 0.57 to 0.94. These encouraging results can guide further development of models to allow for more reliable ICU mortality predictions.

This paper is a joint WIUT-UoW Research Collaboration project coauthored by Nodira Nazyrova, Ikboljon Sobirov and Dr. Abdumalik Djumanov (Computing Department, WIUT) and Mahmoud Aldraimli and Prof. Thierry Chaussalet (the School of Computer Science and Engineering and the Health Innovation Ecosystem (HIE) at the University of Westminster, London). Recently the paper was presented at the International Symposium on Bioinformatics and Biomedicine (BioInfoMed'2020) and received the Best Oral Presentation Award.

BIO: Nodira Nazyrova has been joining WIUT since 2015, and currently is the module leader for ‘Machine Learning and Data Analytics’ and ‘Business Intelligence’ at Computing Department. Her research interests focus on Machine Learning. Nodira Nazyrova is WIUT alumnus (BSc Business Information Systems, Class of 2012). She received her Master of Science degree in Computer Science from TH Koln, Germany.

Research Seminars are a place where faculty and research assistants share completed research and work in progress. They receive valuable feedback to strengthen their research preparing them for conference presentations and publications. On occasion, visiting researchers from outside the University share their work. Research Seminars are open to public.

Download attached Presentation

Related Articles


Alumni             Virtual Tour
Intranet           FAQ
Web Mail         Students

Quick Links

About Us         Parents        Scientific Council
Careers            News            Silk Road
Research                                Virtual Reception