AUXILIARY RECOMMENDATION MODEL OF TRADITIONAL CHINESE MEDICINE PRESCRIPTIONS BASED ON GROUP COLLABORATIVE FILTERING

In the process of formulating traditional Chinese medicines, doctors must master complex relationships between symptoms and syndromes, medicinal properties and compatibility of herbs, etc., and the efficacy of prescriptions is greatly affected by subjective experience. Therefore, an auxiliary recommendation model of traditional Chinese medicine prescription based on group collaborative filtering is proposed. The co-occurrence factor and weighted preference estimation were added, and the correlation between symptoms and herbs, the collaborative relationship between symptoms were extracted from the implicit information of prescriptions, then the group embedding representation of symptom group was generated by aggregating member preferences, which was used as an input to the model and further learn the preference relationship between symptom groups and herbs, so as to obtain recommended Chinese medicine prescriptions. Experiments on real dataset show that our model outperforms the comparison method in terms of precision, recall, and F1-score.