Dataset from the Depression Community on Weibo, 2018-2024

The study developed a depression risk annotation scheme for social networks and constructed a depression risk prediction model based on deep learning algorithms to dynamically identify depression risks in social networks. The data for the study was sourced from 451,071 publicly available texts obtained through the keyword "depression supertopic" between January 1, 2018, and May 1, 2024. As the study focused solely on user statements within the depression community, texts not originating from the "depression supertopic" were excluded, resulting in an effective dataset containing 344,817 texts from 55,618 community members, with an average of 6.199 texts per member. The study respects the legal rights of participants, and the data consists entirely of publicly available user text data acquired through the Weibo open platform, with all information that could be used to trace members strictly removed or concealed.