Reproducibility data package for SC-DHG: semantic-conditioned dual-channel hypergraph learning for herb-disease association prediction

This dataset provides the reproducibility data package for SC-DHG, a semantic-conditioned dual-channel hypergraph learning framework for herb-disease association prediction. The package contains the relation files, preset negative samples, entity name tables and precomputed biomedical semantic embeddings required to reproduce the experiments on the TCM-suite and Ethnobotany datasets. Generated caches, checkpoints, logs and model outputs are not included. The accompanying source code is available in the corresponding GitHub repository.