A Text Mining and Machine Learning Study on the Trends of and Dynamics between Mental Health and Collective Action

Dr Chan, Christian Shaunlyn   (Principal Investigator (PI))

Mr Lam Calvin   (Co-Investigator)

7

2020-03-19

376642

A Text Mining and Machine Learning Study on the Trends of and Dynamics between Mental Health and Collective Action

Collective Action, Machine Learning, Mental Health, Social Media, Text Mining

Others - relating to Social Sciences

Humanities & Social Sciences (H)

SR2020.A8.017

Public Policy Research Funding Scheme

2020

Completed

The current study seeks to investigate the dynamic relationship between mental health and the 2019 social movement in Hong Kong through the analysis of Cantonese social media platforms, including LIHKG. Using machine learning techniques, the current study will establish a Cantonese Corpus that specifically contains the relevant Cantonese and netizen terminologies and lexicon. We will focus on words related to mood disorder symptoms, sleep symptoms, and collection actions related to the protests against the controversial 2019 Extradition Bill. The established Cantonese Corpus will be used for analyzing the trends of lexicon during the investigation period (June 2019 onward), and also their association with the other terms and words of interest (i.e., those related to mental health symptoms and collective action). The specific objectives are to: (1) establish a comprehensive Cantonese Corpus for text mining social media platforms in Hong Kong; (2) identify the lexicon related to mental health symptoms (especially mood and sleep disorders) and collective action; (3) analyze trends, timelines, and frequencies of using the identified terminologies; (4) analyze the influences of mood and sleep symptoms on collective action and vice versa, through text mining; and (5) derive theoretical and policy implications according to the findings.