基于人工神经网络的藏药方剂配伍规律的研究

ibetan medicine prescriptions are mostly compound prescriptions with complex components, which are caused by the fusion of various drugs. Aiming at the complicated non-linear problem between Tibetan medicine prescriptions and therapeutic effects, this study uses the good nonlinear fitting ability of BP Network to construct a BP Network model, and explore the relationship between the characteristics of Tibetan medicine prescriptions and therapeutic effects, and clarify the internal mechanism and The principle of formula is a systematic study on the compatibility of Tibetan medicine. At the same time, the Deep Belief Network was introduced, and the network was pre-trained with a layer-by-layer training strategy to solve the problems of easy fall into local minimums and disappearance of multiple hidden layers in the BP Network. Studies have shown that the deep belief network can autonomously learn the characteristics of prescriptions, have better predictive ability for the efficacy of Tibetan medicine, and create a new research platform for the study of the compatibility of Tibetan medicine.