A Knowledge Graph-Driven Hypergeometric Efficacy Prediction Model for Classical Traditional Chinese Herbal Formulas

TCM efficacy computational representation: The multilevel semantic structure of traditional Chinese medicine formulas makes their efficacies difficult to represent computationally.

TCM efficacy prediction: There is a need to quantitatively predict the dominant efficacies of classical TCM herbal formulas.

TCM efficacy semantic inconsistency: Traditional efficacy descriptions suffer from semantic inconsistency and incompleteness that limit computability.

TCM efficacy computational representation: The multilevel semantic structure of traditional Chinese medicine formulas makes their efficacies difficult to represent computationally.

TCM efficacy prediction: There is a need to quantitatively predict the dominant efficacies of classical TCM herbal formulas.

TCM semantic knowledge graph: The study constructs a knowledge graph over disease, syndrome, symptom, efficacy, and herb entities to standardize and infer multilevel efficacy relationships.

Hypergeometric Efficacy Prediction Model: The study proposes HEPM, which uses hypergeometric enrichment analysis to test whether efficacies are significantly aggregated within a formula.

Curated classical formula validation dataset: The model is validated on a curated dataset of 174 classical formulas from authoritative TCM sources.

TCM semantic knowledge graph: The study constructs a knowledge graph over disease, syndrome, symptom, efficacy, and herb entities to standardize and infer multilevel efficacy relationships.

Hypergeometric Efficacy Prediction Model: The study proposes HEPM, which uses hypergeometric enrichment analysis to test whether efficacies are significantly aggregated within a formula.

Characteristic efficacy reproduction:

F1 score 0.63:

Improved efficacy computability: