机器学习在心肌梗塞并发症中的预测研究

The Prediction of Myocardial Infarction Complications with Machine Learning

为了降低心肌梗塞患者伴有并发症的发生率,利用机器学习方法构建心肌梗塞并发症预测模型,以心肌梗塞患者的医疗数据作为输入,以心肌梗塞患者的并发症类型作为输出,辅助临床医务人员早期判断,提前采取必要的干预措施。研究结果表明,Linear_SVM模型的整体预测性能优于MLP模型和RBF_SVM模型,其预测准确率为76.28%,特别是在心房纤颤、三度房室传导阻滞、心肌破裂和心肌梗死后综合征上表现出较好的预测效果。

In order to reduce the incidence of myocardial infarction patients accompanied by complications,using machine learning method to build prediction models of myocardial infarction complication in patients with myocardial infarction medical data as input,with myocardial infarction patients with complications of type as the output,auxiliary judgment early clinical medical personnel,take the necessary intervention measures in advance.The results show that the overall prediction performance of Linear_SVM model is better than that of MLP model and RBF_SVM model,with a prediction accuracy of 76.28%,especially for atrial fibrillation,third-degree atioventricular block,myocardial rupture and post-myocardial infarction syndrome.