This repo contains the experimental codes for the paper "Psychiatric Scale Guided Risky Post Screening for Early Detection of Depression". (IJCAI 2022 AI for Good).
The 3 folders contains the code for each dataset. Detailed information an be seen in the readme inside the directories.
The main experiments are conducted on erisk2017 dataset, while we also test the generalizability of the proposed model on the other 2 datasets (See Appendix).
We can't include the dataset here due to the terms and restrictions of them, but their their information can be found in eRisk2017, RSDD and Topic-Restircted (Retrieved with the implementation in https://github.com/kharrigian/emnlp-2020-mental-health-generalization)