Accessibility navigation


Machine learning based biomarkers for neurodegenerative disease classification

Varzandian, A. (2023) Machine learning based biomarkers for neurodegenerative disease classification. PhD thesis, University of Reading

[img]
Preview
Text - Thesis
· Please see our End User Agreement before downloading.

2MB
[img] Text - Thesis Deposit Form
· Restricted to Repository staff only

4MB

It is advisable to refer to the publisher's version if you intend to cite from this work. See Guidance on citing.

To link to this item DOI: 10.48683/1926.00111984

Abstract/Summary

In this thesis a novel classification model framework to predict Alzheimer’s Disease is described. In this work a novel brain age feature is proposed, which estimates the biological age of parts of the brain affected by Alzheimer’s Disease. This feature can act as a biomarker for medical professionals which together with age, can make an Alzheimer’s Disease prediction with high performance. In addition to this feature, a novel interpretable classification framework is proposed for prediction of AD which can achieve high classification performance. Also, a novel interpretability index is also proposed which indicates to the medical professionals why such prediction has been made and which input features had the greatest impact on the final output. The brain age Alzheimer’s Disease prediction model is also applied to other type and stages of dementia in a multi-class classification setting as an extension of the work. The results achieved in this thesis in both binary and multi-class classification are comparable to the baseline and relevant previous literature. The binary classification accuracy achieved are 92.84% and 89.74% for female and male subjects respectively.

Item Type:Thesis (PhD)
Thesis Supervisor:Di Fatta, G.
Thesis/Report Department:Department of Computer Science and Engineering
Identification Number/DOI:https://doi.org/10.48683/1926.00111984
Divisions:Science > School of Mathematical, Physical and Computational Sciences > Department of Computer Science
ID Code:111984
Date on Title Page:October 2022

Downloads

Downloads per month over past year

University Staff: Request a correction | Centaur Editors: Update this record

Page navigation