目的:探讨决策树与神经网络模型在预测精神分裂症患者复发中的应用,为精神分裂症的防治提供参考。方法:通过对首发住院精神分裂症患者两年的随访调查,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76.66%和82.73%,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首发住院精神分裂症患者复发的预测是可行的。%Objective:To explore decision tree (DT) and Artificial Neural Networks (ANN) models to predict the relapse of the First-episode Schizophrenic Patients,and to provide a reference for the prevention and treatment of schizophrenia.Method:Conducted a two-year follow-up survey of the first episode schizophrenic patients in our hospital from 2009 to 2011,and developed C5.0 DT and BP ANN two predictive models.Result:The C5.0 DT model achieved a classification accuracy of 76.66%,and BP ANN model achieved a classification accuracy of 82.73%. The BP ANN model performed better than C5.0 DT model. Conclusion:Using DT and ANN models to prediction the relapse of first episode schizophrenic patients is feasible.
展开▼