NORMA eResearch @NCI Library

Diabetes Diagnosis and Readmission Risks Predictive Modelling: USA

Reid, Clodagh (2019) Diabetes Diagnosis and Readmission Risks Predictive Modelling: USA. Masters thesis, Dublin, National College of Ireland.

[img]
Preview
PDF (Master of Science)
Download (1MB) | Preview

Abstract

One of the most critical healthcare problems today is diabetes. In the US, 30 million people are affected by diabetes with healthcare related costs at a sobering $327 billion annually. Long term effects if untreated can result in damage to the heart, blood vessels, eyes, kidneys, feet, nerves and even mortality from heart attack or stroke. As diabetic patients have increased, consequently the number of diabetic hospital readmissions have become greater. Early readmission can impact patient health, operational efficiency and cost burden. The aim of this research is to examine diabetes diagnosis and early readmission with the same dataset. A comprehensive methodology with pre-processing and transformation which comprised of permutation feature importance, feature engineering and SMOTE are some of the methods used to deal with noisy, inconsistent, imbalanced data. Predictive models include LR, BDT, SVM, NN, DF. The best performing model is selected to create a web service where users can input data and receive scored results connecting the user to the data. Metrics include accuracy, recall and AUC to measure the performance of the models where a Boosted Decision Tree achieved the highest results. Hospital readmission accuracy at 86% and diabetes diagnosis accuracy at 67%.

Item Type: Thesis (Masters)
Subjects: Q Science > QA Mathematics > Electronic computers. Computer science
T Technology > T Technology (General) > Information Technology > Electronic computers. Computer science

Q Science > QA Mathematics > Computer software
T Technology > T Technology (General) > Information Technology > Computer software

R Medicine > Healthcare Industry
Divisions: School of Computing > Master of Science in Data Analytics
Depositing User: CAOIMHE NI MHAICIN
Date Deposited: 27 Nov 2019 12:18
Last Modified: 27 Nov 2019 12:18
URI: http://trap.ncirl.ie/id/eprint/4106

Actions (login required)

View Item View Item