Information Needs Mining of COVID-19 in Chinese Online Health Communities

Wang, Jie; Wang, Lei; Xu, Jing; Peng, Yan

This study explores the information needs for the novel coronavirus pneumonia (COVID-19) in Chinese online health communities (OHCs). Based on the question and answer data about COVID-19 in six Chinese OHCs, topic mining and data analysis were conducted. We propose a CL-LDA topic model (Latent Dirichlet Allocation Model with co-occurrence of lexical meaning) based on lexical meaning co-occurrence analysis and LDA topic model. Four main information need topics and their proportion are found in this study, including symptom (45.50%), prevention (36.11%), inspection (10.97%), and treatment (7.42%). We also discover that men are most concerned about symptom information while women are most concerned about prevention information; young users have the largest proportion of information needs, and they are most concerned about prevention information. Experiment results show that the CL-LDA model can well adapt to the topic mining task of short text which is semantic sparse and lacking co-occurrence information in OHCs. The research results are helpful for OHCs to provide accurate information assistance and improve service quality.