ABSTRACT
Aiming at the serious harm of depression, a depression intervention model from the perspective of computational social science is proposed. The proposed model is based on the social media data of users. The LSTM model was used to analyze the potential depressive tendencies of users, while the psychiatric PHQ-9 depression scale was employed as a tool to diagnose depression and measure depression severity. The LDA model was utilized to identify the topics that users are concerned about, and these topics were considered important sources of their depressive moods. Subsequently, corresponding book prescriptions were provided for the implementation of bibliotherapy to carry out sociological intervention. A quantitative evaluation system for a depression intervention model from the perspective of computational social science was established to an extent in the present study, which promotes the standardization of depression interventions. This system would significantly reduce the consumption of health resources, with real social and economic benefits.
