The invention discloses a
drug specification-based reasonable medication
knowledge graph construction method, which comprises the following steps of S10, extracting a
drug specification, and inducingentities and relationships in the
drug specification through an expert labeling method to form an entity and relationship indexing rule base; S20, training a
machine learning model through a semi-
supervised learning method on the basis of data labeled by experts and
machine learning rules; and S30, using the trained
machine learning model to predict and
label the unlabeled drug specification to form a
knowledge graph of the drug relationship. According to the method,
natural language processing and
knowledge graph technologies are utilized, and the reasonable medication
knowledge base is constructed based on the
medicine specification, so that medication errors can be avoided, medical risks can be prevented, and the clinical medication safety can be improved. The method has an automatic construction capability, and the construction precision is improved through
machine learning.