The invention discloses a
drug dose prediction method based on patient similarity, and belongs to the technical field of
drug dose prediction, and the method comprises the steps: obtaining clinical real historical data of historical patients, and carrying out the preprocessing and coding of the data, and obtaining the representation data of the historical patients; acquiring data of a current patient, and preprocessing and encoding the data to obtain characterization data of the current patient; according to the characterization data of the historical patient and the characterization data of the current patient, a multi-feature similarity measurement strategy, feature
correlation analysis and a
machine learning model are used for calculation and
processing, and a medication dosage prediction result of the current patient is obtained. According to the method, standardized
processing and analysis are carried out on data of patients, the most similar
patient group is matched for each patient from the aspects of characteristics and patient integration, and the medication dosage of the patient is predicted by adopting a
machine learning model, so that the medical efficiency can be improved, the problems of poor consistency and poor
interpretability of the medication dosage in clinic can be solved, and the clinical application prospect is wide. Medical resources are effectively utilized, and waste is reduced.