A Novel TCM Prescription Recommendation Algorithm Based on Deep Crossing Neural Network

In this manuscript, a TCM prescription recommendation model based on deep crossed neural network was proposed, which aims to solve the problem caused by sparse discrete features in TCM prescription information obtained after encoding and feature automatic cross-combination through embedding layer and residual network. Use residual neural network to reduce overfitting and make convergence faster. Using the multi-level residual network, the combined feature vector of medical records and prescriptions is fully residual operation, so that the model can obtain more nonlinear feature information between medical records and prescriptions, and realize the corresponding treatment according to the name of the disease. The function of prescription to improve the level of clinical treatment.