TCM_ICPR: Personalized Recommendation Algorithm for Traditional Chinese Medicine Prescriptions Based on Knowledge Graph Enhancement

With the development of artificial intelligence (AI) and machine learning (ML) in Traditional Chinese medicine (TCM) clinical diagnosis and treatment, the existing TCM prescription recommendation model lacks the personalized characteristics of patients, which does not conform to the process of clinical TCM diagnosis and brings a great obstacle to the progress of TCM. To remedy this deficiency, we propose an interpretable algorithmic model for TCM prescription recommendation (TCMICPR). We utilize graph neural network (GNN) and Multilayer Perceptron (MLP) to learn symptoms and herb representations. To better simulate the clinical diagnosis process, we leverage the attention mechanism to learn the weights of symptoms. In addition, to improve the accuracy of the model, we manually constructed a large-scale TCM knowledge graph as a supplement of additional knowledge, which covers most of the knowledge related to herb and symptoms, etc. Experimental results on two public datasets show that our model has a higher accuracy.