摘要
目的 探讨决策树与神经网络模型在预测颅脑损伤患者预后中的应用.方法 通过对重型颅脑损伤患者入院情况的分析并结合预后随访结果,采用C5.0决策树与BP神经网络方法分别建立预测模型.结果 两种预测模型的预测准确度分别为75.39％和82.20％,神经网络方法建立的预测模型要优于决策树模型.结论 采用决策树与神经网络模型来建立对颅脑损伤患者的预后预测是可行的.
Objective To investigate the application of decision tree （DT） and artificial neural networks （ANN）in predicting prognosis for patients with traumatic brain injury.Methods According to the admission information and follow-up results,C5.0 DT and back-propagation artificial neural networks （BP-ANN） were applied to establish the prognostic models.Results The C5.0 DT model achieved a classification accuracy of 75.39％,and the BP-ANN model achieved a classification accuracy of 82.20％.The BP-ANN model performed better than C5.0 DT model.Conclusion Using DT and ANN models in predicting the prognosis of traumatic brain injury patients is feasible.
出处
《临床外科杂志》
2013年第9期694-696,共3页
Journal of Clinical Surgery
关键词
数据挖掘
预测
颅脑损伤
data mining
prediction
brain injury
