Lung postoperative complication risk value prediction method, device, equipment and medium

The invention discloses a lung postoperative complication risk value prediction method and device, equipment and a medium. The method aims at solving the problems that the modeling capacity of a neural network in an existing prediction model for heterogeneous tables is poor, so that the model performance is affected, and the risk of lung postoperative complications cannot be fully reflected due to the fact that data sources of the prediction model are not comprehensive. Whether complications occur in the lung postoperative sample data or not is calibrated, and label data is formed; constructing a risk value prediction model, wherein the risk value prediction model comprises an XGBoost tree module, an LSTM module and an nmODE module; training the risk value prediction model by adopting the sample data; and acquiring real-time postoperative data, inputting the real-time postoperative data into the risk value prediction model, and outputting a predicted risk value by the risk value prediction model. According to