[Retracted] A Random Walk with Restart Model Based on Common Neighbors for Predicting the Clinical Drug Combinations on Coronary Heart Disease

Table 2

Work summary.

Evaluation tools	MATLAB
Performance metrics	SE, ROCs’ curve, and values of AUC
Case studies	The clinical drug combinations on coronary heart disease
Deployment strategy	Data management of the outpatient prescription: data collection, data preprocessing
	Model learning: design CN-RWR model, model selection (RWR), training, parameter selection (c in leave-one-out cross-validation (LOOCV))
	Model verification: simulation-based testing
Advantages	The prediction algorithm in this study was based on the topological properties of a drug combinations network in the real world and it makes the predicted results more similar to the drug combinations of the real world; Our model performance is better than the traditional one
Disadvantages	The predictive performance of the model can be further improved