University of Illinois at Chicago
Browse

File(s) under embargo

1

year(s)

2

month(s)

7

day(s)

until file(s) become available

GraDD: A Graph Machine Learning and Natural Language Processing Approach to Automatic Dementia Detection

thesis
posted on 2023-05-01, 00:00 authored by Edoardo Stoppa
Alzheimer's Disease is a progressive neurodegenerative disease that has no cure, with a lengthy and costly diagnosis process. One of the earliest signs of Dementia is difficulty in producing language, and that is why Computer-Aided Dementia Detection through Natural Language Processing is emerging as a viable solution for an early diagnosis. Our work makes three main contributions. First, we have developed a framework used for the extraction of manually engineered features from dementia-related audio files and transcripts. The framework is open-source and easily extensible, making it possible for other researchers to contribute by adding new features or new groups of features. Second, we have built a machine learning performance evaluation framework that standardizes the entire performance analysis process. Last, we propose a new approach to Automatic Dementia Detection, based on Graph Machine Learning. We demonstrate that the method is promising given the encouraging results, even if it's still far from the current state-of-the-art performance.

History

Advisor

Parde, Natalie

Chair

Parde, Natalie

Department

Computer Science

Degree Grantor

University of Illinois at Chicago

Degree Level

  • Masters

Degree name

MS, Master of Science

Committee Member

Medya, Sourav Leow, Alex Santambrogio, Marco D.

Submitted date

May 2023

Thesis type

application/pdf

Language

  • en

Usage metrics

    Categories

    No categories selected

    Exports

    RefWorks
    BibTeX
    Ref. manager
    Endnote
    DataCite
    NLM
    DC