OSW

SIGNATURE WORK
CONFERENCE & EXHIBITION 2023

Trend of Depression in Online Mutual Help Group for Major Depressive Disorder by a Two-step NLP Method

Name

Wanqi Hu

Major

Data Science

Class

2023

About

Wanqi Hu; Class of 2023 at DKU; Data Science

Signature Work Project Overview

Major depressive disorder (MDD) has become a common problem in recent years. Online mutual help groups provide an interactive channel for patients with MDD to communicate, find support, and pursue a sense of belonging. To quantify the impact of mutual help groups on MDD patients, we introduce a comprehensive two-step Natural Language Processing (NLP) method which allows us to measure the degree of depression and its evolution over group interactions. This method combines a Long Short-Term Memory model and dictionary methods, and it overcomes the limitation of traditional dictionary methods that fail to detect the overall sentiment in cases of complex contents. Based on the data from an MDD mutual help group on the Chinese social platform Douban, the results show that replies are more positive compared with posts in the group, and group members tend to reply to more depressive posts. Moreover, through the within-group interaction, posters become less depressive as they post more in the group.

Signature Work Presentation Video