Phd

With the widespread adoption of social media, an increasing number of people are sharing information related to their lives, emotions, and health on these platforms. This abundance of data provides rich material for studying the relationship between social media use and mental health. By analyzing user behavior and content on social media, we can better understand the early signs of mental health issues, changes in social emotions, and the potential role of social media in mental health interventions.

Paper One: Social Media as an Emotional Barometer: Bidirectional Encoder Representations From Transformers–Long Short-Term Memory Sentiment Analysis on the Evolution of Public Sentiments During Influenza A on Sina Weibo

This paper examined the evolution of public emotions on Sina Weibo during the 2023–2024 Influenza A outbreak using a BERT-LSTM deep learning sentiment analysis model. By analyzing more than 53,000 Weibo posts, the study classified emotions into six categories: happiness, neutrality, sadness, fear, anger, and surprise. The findings revealed that sadness was the dominant emotional response throughout the outbreak and remained elevated even after infection rates declined, indicating the long-term psychological impact of public health crises. The study further demonstrated that emotional dynamics on social media closely corresponded with changes in influenza positivity rates and broader social contexts such as holidays and population movement. Grounded in Social Sharing of Emotion theory, the research highlighted how social media functions as an “emotional barometer” capable of capturing collective psychological fluctuations in real time, offering valuable implications for digital mental health monitoring and public health communication.

Published DOI: http://10.2196/68205  

Paper Two: Temporal Dynamics of Influenza-Associated Anxiety Symptom Linguistic Markers on Weibo (2023–2024): A Study of Incidence Correlation, Diurnal Patterns, and Time-Series Trends 

This paper investigated anxiety-related expressions on Sina Weibo during the 2023–2024 Influenza A outbreak by analyzing large-scale social media data based on DSM-IV anxiety symptom dimensions. Using natural language processing and time-series analysis, the study explored how different anxiety symptoms evolved across outbreak stages and daily time periods, as well as their association with influenza incidence rates. The findings revealed significant temporal fluctuations in anxiety-related linguistic markers, with symptoms such as fatigue and irritability remaining particularly prominent throughout the outbreak. The study demonstrated that social media can function as a real-time digital surveillance tool for monitoring collective anxiety dynamics and public mental health during public health crises.

Published DOI: http://10.2196/preprints.88849

Paper Three: Social Support Structures Following Anxiety Expression on Sina Weibo During the Influenza A Outbreak

Building upon the first two studies, the third paper further explores the relationship between social media use and mental health by examining how anxiety-related disclosures on Sina Weibo during the 2023–2024 Influenza A outbreak triggered different forms of social support within comment sections. Grounded in Social Sharing of Emotion Theory and Functional Social Support Theory, the study analyzes how users engage in emotional co-regulation, informational coping, and online interaction during public health crises.

Using large-scale Weibo post–comment data and a multi-label deep learning model, the study identifies different support structures, including emotional, informational, instrumental, and ineffective support. It further examines variations across anxiety symptom types, gender, and outbreak stages. The findings highlight that social media functions not only as a space for emotional expression but also as a platform for collective psychological regulation and digital social support during health emergencies. 

Research Methods

AI and Natural Language Processing (NLP) Techniques: Sentiment analysis, topic modeling, and keyword extraction are used to analyze textual data on social media.

• Gain a deep understanding of how social media reflects and influences mental health.
• Propose social media-based strategies for mental health monitoring and intervention.
• Provide new insights and methods for research and practice in the fields of public health and mental health.