Novel Machine Learning Tools to Enhance Mortality Prediction

Simple item page
dc.contributor.advisorYesha, Yelena
dc.contributor.authorMativo, Isaac
dc.contributor.departmentComputer Science and Electrical Engineering
dc.contributor.programComputer Science
dc.date.accessioned2022-02-09T15:52:34Z
dc.date.available2022-02-09T15:52:34Z
dc.date.issued2020-01-01
dc.description.abstractThe tremendous growth of clinical data, both in healthcare institutions and personal devices, has led to the development of smart systems that can analyze this data to understand patterns and make predictions. The goal of understanding this data is to help make informed clinical decisions that improve health outcomes and save costs. Patient data has been leveraged by these intelligent systems to notice features that are important and actionable in improving clinical outcomes. One of the key clinical outcomes especially for intensive care unit (ICU) patients is mortality. Mortality prediction for ICU patients continues to be an important challenge in clinical care because of the complex nature of patients and data involved. With accurate mortality prediction, patients can be classified into different risk categories to ensure proper care is given therefore allowing for the best possible clinical outcomes and cost savings. Several mortality prediction models have been developed some of which use Machine Learning techniques. However, the existing mortality prediction tools are not generalizable to entire patient populations and are not designed to predict a diverse range of clinical endpoints. Additionally, these mortality prediction models are not personalized to the patients' unique and changing clinical profile. In this dissertations, we present a mortality prediction approach that addresses these challenges thereby allowing for generalizable, extensible, and personalized mortality prediction modeling. We achieve this goal while improving prediction accuracy over existing state of the art tools. We present a novel methodology of combining clinical biomarker data, patient similarity features, and existing prediction tool output to predict ICU mortality. We examine patient comorbidities to discover their impact on mortality and use them as attributes in our modeling. We capture biomarker features as used in a severity scoring tool as in integral input in our modeling. In so doing, we have designed an approach to clinical prediction that uses matured machine learning techniques to combine patient similarity measures and patient biomarker information to improve prediction accuracy. Our approach builds on previous work that has been done in clinical outcome prediction and lays a strong foundation for continued innovation.
dc.formatapplication:pdf
dc.genredissertations
dc.identifierdoi:10.13016/m2x8ew-vdt8
dc.identifier.other12391
dc.identifier.urihttp://hdl.handle.net/11603/24177
dc.languageen
dc.relation.isAvailableAtThe University of Maryland, Baltimore County (UMBC)
dc.relation.ispartofUMBC Computer Science and Electrical Engineering Department Collection
dc.relation.ispartofUMBC Theses and Dissertations Collection
dc.relation.ispartofUMBC Graduate School Collection
dc.relation.ispartofUMBC Student Collection
dc.sourceOriginal File Name: Mativo_umbc_0434D_12391.pdf
dc.titleNovel Machine Learning Tools to Enhance Mortality Prediction
dc.typeText
dcterms.accessRightsAccess limited to the UMBC community. Item may possibly be obtained via Interlibrary Loan through a local library, pending author/copyright holder's permission.
dcterms.accessRightsThis item may be protected under Title 17 of the U.S. Copyright Law. It is made available by UMBC for non-commercial research and education. For permission to publish or reproduce, please see http://aok.lib.umbc.edu/specoll/repro.php or contact Special Collections at speccoll(at)umbc.edu

Files

Original bundle
Now showing 1 - 1 of 1
No Thumbnail Available
Name:
Mativo_umbc_0434D_12391.pdf
Size:
1.62 MB
Format:
Adobe Portable Document Format
Download

Collections

UMBC Theses and Dissertations
UMBC Computer Science and Electrical Engineering Department
UMBC Graduate School
UMBC Student Collection