A sentiment analysis and topic modeling of weibo discourse on public perceptions of breast cancer

Objective: Social media has emerged as a transformative platform for data-driven health communication about breast cancer. To reveal social cognition and enhance public awareness of breast cancer, this study analyzes emotional tendencies and core topics in Weibo discourse. Methods: A web crawler was employed to collect 185,486 Weibo posts containing the keyword "breast cancer", with 134,588 posts retained after preprocessing. The discourse analysis involved, first, a lexicon-based sentiment classification of the text data, and subsequently, the application of Latent Dirichlet Allocation (LDA) topic modeling of each identified sentiment. Results: This study identified three dominant emotional categories-positive (40.5%), neutral (34.5%), and negative (25%)-along with distinct thematic patterns associated with each. Salient topics included "personal experiences" and "symptoms," with positive posts emphasizing recovery narratives, clinical advances, and prevention; neutral discourse focusing on health management and support systems; and negative discourse centered on emotional distress, treatment challenges, and risk factors. Discourse analysis reveals the significance of emotional support, access to medical and insurance resources, and reliable scientific information in fighting breast cancer. Furthermore, breast cancer is discursively constructed through three key frames such as "a women's issue", "a public health crisis", and an issue of "individual responsibility". Various stakeholders-including media, government, medical institutions, and patient groups-played distinct roles in the discursive construction of breast cancer. Conclusion: Our findings provide critical insights into public needs on breast cancer, informing the development of targeted strategies to enhance health awareness, shape attitudes, and improve stakeholder-led education and advocacy.