论文部分内容阅读
为提高医疗决策的效率和有效性,建立了邻域粗糙集融合贝叶斯神经网络的组合医疗决策模型。首先,通过邻域粗糙集对医疗决策系统进行知识约简,去除系统中冗余、干扰的指标,提取关键指标,并将约简后的指标作为神经网络模型的输入指标;之后针对BP神经网络容易过拟合的缺点,采用贝叶斯正则化方法对神经网络进行优化,提高输出指标的预测效果和效率。最后,通过一个心脏病医疗诊断实例对模型的应用实现效果加以分析及验证,结果表明,模型测试集分类准确率达到了88.89%;并将实验结果同几种常见的属性选择方法、分类模型以及2008~2014年的8个历史参考文献进行了对比。
To improve the efficiency and effectiveness of medical decision-making, a combined medical decision model of neighborhood rough set fusion Bayesian neural network is established. Firstly, the knowledge reduction of medical decision-making system is carried out by using the neighborhood rough set to remove the redundant and interference indicators in the system, and the key indicators are extracted, and the reduced indicators are taken as the input index of the neural network model. Then, Easy over-fitting shortcomings, the use of Bayesian regularization method to optimize the neural network to improve the output index of the prediction and efficiency. Finally, a case of heart disease diagnosis was used to analyze and verify the application of the model. The results showed that the classification accuracy of the model test set reached 88.89%. The experimental results were compared with several common attribute selection methods, classification models and Eight historical references from 2008 to 2014 are contrasted.