Development of Input-Output Hidden Markov Model for Estimating Diabetes Progression

Tianbei Zhang

2022

Abstract

Incessant unhealthy routines are believed to induce chronic diseases. However, current modelling can barely estimate diabetes progression by analyzing daily behaviours. In this study, an input-output hidden markov model (iohmm) was constructed to forecast the progression of diabetes mellitus based on ordinary routines and to reveal the association among illness indicators (e.g., blood glucose levels), living habits and medical interventions. The analysis of diabetes datasets from the ucirvine machine learning repository revealed that the high amount of food intake, insulin overdose and unideal health status could increase the risk of severe exacerbation in diabetes patients. It was also found that the variation of blood glucose increased as the patients’ health conditions worsened. Besides, among all the factors tested in this study, the patients’ initial health conditions contributed the most to blood sugar fluctuation, while minor contributions from meal and insulin were still effective enough to be regarded as significant factors. The proposed iohmm model enables the inference of patient’s health conditions by analyzing their living habits. Most importantly, this study successfully developed a novel iohmm model to estimate diabetes progression, which can be generalized and applied to other chronic diseases.

Paper Citation

in Harvard Style

Zhang T. (2022). Development of Input-Output Hidden Markov Model for Estimating Diabetes Progression. In Proceedings of the 4th International Conference on Biotechnology and Biomedicine - Volume 1: ICBB; ISBN 978-989-758-637-8, SciTePress, pages 215-222. DOI: 10.5220/0012018600003633

in Bibtex Style

@conference{icbb22,
author={Tianbei Zhang},
title={Development of Input-Output Hidden Markov Model for Estimating Diabetes Progression},
booktitle={Proceedings of the 4th International Conference on Biotechnology and Biomedicine - Volume 1: ICBB},
year={2022},
pages={215-222},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0012018600003633},
isbn={978-989-758-637-8},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 4th International Conference on Biotechnology and Biomedicine - Volume 1: ICBB
TI - Development of Input-Output Hidden Markov Model for Estimating Diabetes Progression
SN - 978-989-758-637-8
AU - Zhang T.
PY - 2022
SP - 215
EP - 222
DO - 10.5220/0012018600003633
PB - SciTePress