Method for predicting death probability of hospitalized patient with coronary heart disease

The invention discloses a method for predicting the death probability of a coronary heart disease inpatient. The method comprises the following steps: step 1, defining a target; step 2, data acquisition; step 3, constructing a prediction model; 4, evaluating a model effect; 5, verifying and updating the model; in the step 2, data required by model construction comprises information of patients, information of treatment, information of departments and information of doctors; in the step 4, the evaluation index is an F1 value or an AUC index; according to the method, the machine learning technology is utilized to analyze the full-dimensional information of the inpatient with the coronary heart disease, and the death probability of the specific patient is predicted, so that an auxiliary decision making basis is provided for clinical judgment, the accuracy and timeliness of the clinical judgment are improved, a doctor and the patient can make a joint decision better by means of the prediction model, and the predi