Predicted Efficacy Results for 174 Classical TCM Formulas (Using Hypergeometric Distribution Model)

This dataset was developed as part of the major research project under the Innovation Program of the China Academy of Chinese Medical Sciences—Study on the Decomposition-Based Efficacy Prediction Model for Chinese Formulas under Big Data Context (Project No. CI2021A00508). The study constructed a hypergeometric distribution-based prediction model for herbal formula efficacy, using 174 classical TCM formulas as the sample set.Efficacy prediction was performed using the hypergeometric probability function, and model performance was evaluated using the F1-score metric. The evaluation was based on a comparison between the model-predicted efficacies and the originally documented efficacies of the 174 formulas. The resulting F1-score was 0.63, indicating a good level of predictive performance.This dataset provides the complete set of input variables and predicted results for each of the 174 formulas, as computed by the hypergeometric model.n — Number of herbs in the formula: The number of herbs in the predicted formula (e.g., There are four herbs in predicted formula. So n=4).N — Total number of herbs in the domain: The total number of herbs within the domain (e.g., the knowledge graph contains around 400 herbs, So N=400).k — Number of herbs in the formula with a specific efficacy: The count of herbs in the predicted formula that contribute to a specific efficacy (e.g. There are 2 herbs in predicted formula related to "clearing heat" .So k=2 ).M — Total number of herbs in the domain with a specific efficacy: The count of herbs associated with a specific efficacy across all herbs in the domain (e.g. There are 20 herbs associated with "clearing heat" in the domain. So M=20).