Generative algorithms to improve mental health issue detection

Bachelor thesis (2021)

Authors

W.H. Lam Electrical Engineering, Mathematics and Computer Science

Contributors

W.P. Brinkman Interactive Intelligence - EEMCS (mentor)
M. Bruijnes Interactive Intelligence - EEMCS (graduation committee member)
H.S. Hung Pattern Recognition and Bioinformatics - EEMCS (mentor)
Faculty
Electrical Engineering, Mathematics and Computer Science, Electrical Engineering, Mathematics and Computer Science
Copyright: © 2021 Jimmy Lam
More Info
expand_more

Published Date

01-07-2021

Language

English

Reuse Rights

Other than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons.

Faculty

Electrical Engineering, Mathematics and Computer Science

Abstract

Schema therapy is a physiological treatment technique for mental health issues. Based on the thoughts and behaviour, patients are classified to a schema mode which represents their current state of mind. Automatically classifying these thoughts and behaviours could improve detection of potential mental health issues as well as provide better and faster recovery. This research attempts to effectively generate schema-based stories that would be used to train machine learning models such as Support Vector Machines and Recurrent Neural Networks to classify stories from patients about their daily experiences. Experimental evaluation using the OpenAI GPT-2 model shows that it is possible to generate correct and coherent stories with a minimum of 58.7\% correctly classified samples even with sub-optimal data. Using conditional prefixed queries, the OpenAI GPT-2 model can generate stories that resemble the given data but with little to no similarity in terms of BLEU scores.

Files

Research_Project_1.pdf
(pdf | 0.375 Mb)
Unknown license