Study and processing of spoken language of Greek-speaking people with Alzheimer's Disease

Postgraduate Thesis uoadl:3292135 232 Read counter

Unit:
Specialty Language Technology
Πληροφορική
Deposit date:
2023-03-14
Year:
2023
Author:
Valaskatzis Ilias
Supervisors info:
Βαρλοκώστα Σπυριδούλα, Καθηγήτρια Ψυχογλωσσολογίας, Εργαστήριο Ψυχογλωσσολογίας και Νευρογλωσσολογίας, Τομέας Γλωσσολογίας, Τμήμα Φιλολογίας, ΕΚΠΑ.

Σταμούλη Σπυριδούλα, Συνεργαζόμενη ερευνήτρια ΙΕΛ, Διδάκτωρ Γλωσσολογίας ΕΚΠΑ, Ινστιτούτο Επεξεργασίας του Λόγου [ΙΕΛ]/ Ε.Κ. Αθηνά.

Κατσαμάνης Αθανάσιος, Κύριος ερευνητής ΙΕΛ, Γενικός Διευθυντής Τεχνολογίας της Behavioral Signal Technologies Inc, Ινστιτούτο Επεξεργασίας του Λόγου [ΙΕΛ]/ Ε.Κ. Αθηνά.
Original Title:
Μελέτη και επεξεργασία προφορικού λόγου ελληνόφωνων ατόμων με νόσο Alzheimer
Languages:
Greek
Translated title:
Study and processing of spoken language of Greek-speaking people with Alzheimer's Disease
Summary:
In this thesis, which is being carried out within the framework of the Interdisciplinary Postgraduate Programme of Studies in Language Technology (Department of Informatics and Telecommunications, University of Athens), the design of the Corpus of Pronunciation of Greek Speakers with Neurocognitive Disorders [CPGSND] is presented. The thematic core of the project is the collection, processing and analysis of continuous spoken speech data from a Greek-speaking population with mild Alzheimer's Disease. The data collection and processing were carried out according to the specifications of the multilingual repository DementiaBank/TalkBank, as the ultimate goal of the research is to make the data of the experimental group available to this repository. More specifically, a standard protocol for collecting continuous speech data from people with Alzheimer's disease was utilized and the specifications of the standard CHAT system were followed for transcribing and annotating the data. For our experimental group, a total of seven women and ten men, aged fifty to eighty years, were collected, while for the control group, data were collected from twelve women and eight men of the same age range and without neurocognitive disorders [Neurocognitive Disorders]. The labelled data is sought in the future to be used as a training corpus for a classifier to distinguish between individuals with Alzheimer's disease and healthy adults based on their spoken language.
Main subject category:
Technology - Computer science
Keywords:
text corpus, Alzheimer's disease, neurocognitive disorders, continuous speech, Greek speakers, DementiaBank
Index:
Yes
Number of index pages:
4
Contains images:
Yes
Number of references:
31
Number of pages:
41
master_thesis_ Valaskatzis.pdf (986 KB) Open in new window