Readmission prediction model training method, readmission prediction method, device and equipment

The invention provides a re-admission prediction model training method, a re-admission prediction method, a re-admission prediction device and equipment. The method comprises the following steps: acquiring clinical data of a historical heart failure patient; based on the clinical data, extracting features related to the re-admission outcome of the heart failure patient to obtain a first training set; based on the first training set, constructing a causal network model, and training the causal network model to obtain an optimal causal network structure; on the basis of the optimal causal network structure, eliminating features, which have no causal relationship with the re-admission ending of the heart failure patient, in the first training set to obtain a second training set; and based on the second training set, constructing a target re-admission prediction model, and training the target re-admission prediction model to obtain a trained target re-admission prediction model. By extracting the features related