Knowledge graph-based liver cancer postoperative recurrence risk prediction method and device

The invention provides a method and a device for predicting postoperative recurrence risk of liver cancer based on a knowledge graph, and belongs to the field of medical data prediction.The method comprises the following steps: preprocessing patient data to obtain preprocessed data; aiming at the preprocessed data, constructing a knowledge graph containing liver cancer recurrence related influence factors, liver cancer indexes and patient data; an XLNet training model is adopted to train entities and relations in the knowledge graph, and representation vectors of the entities and the relations of the patients are obtained; according to the characterization vector of the patient entity and the relation, an XGB algorithm is adopted for prediction, MSE serves as a loss function, and a corresponding prediction value with the minimum loss function serves as a prediction value of the liver cancer postoperative recurrence risk. The system comprises a data preprocessing module, a knowledge graph construction module,