【摘要】 目的将计算机人工神经网络技术与世界卫生组织（WHO）推出的复合性国际诊断交谈表（CIDI）结合，开发CIDI计算机神经网络诊断系统（CIDI-ANN），以辅助精神疾病的诊断。方法建立误差反向传播网络（BP网络）和自组织映射网络（kohonn网络）辅助CIDI诊断神经症和精神分裂症。使用60例样本训练神经网络，建立CIDI-ANN，使用另外的222例样本进行跨效度检验。结果对WHO提供的CIDI诊断软件CIDI/国际疾病分类第10版、CIDI/精神障碍诊断与统计手册第3版修订本与新建立的CIDI－ANN进行平行测试，3个系统与临床诊断的总符合率的Kappa值分别为0．77，0．79和0．94，对疑难病例的诊断符合率的Kappa值分别为0．48，0．33和0．69。结论在辅助CIDI诊断中，人工神经网络方法优于传统的人工智能方法。人工神经网络与诊断量表的结合是一个新的有前途的研究方向。更多还原

【Abstract】 Objective Artificial Neural Network(ANN), as a potential powerful classifier, was exploredto assist psyhiatric diagnosis of the Composite Intemational Diagnostic Interview(CIDI). Methods Both BackPrpagation(BP) and Kohone networks were developed to fit psychiatric diagnosis and programme (using 60 cases) to classify neurosis, schizophrenia and normal peope. The programmed networks were cross-tested using another 222 cases. All subjects were randomly selected from two psychiatric hospitals in Beijing. Results Compared toICD-10 diagnoois by psychiatrists, the overall kappa of CID/ICD-10. CIDI/DSM-Ⅲ-R and CIDI-ANN were 0.77,0.79 and 0.94. In classifying patients Who were difficult to diagnose, the kappa were 0.48, 0.33 and 0.69 respectively. Conclusions ANN was more powerful than traditional expert system and might be a new method to improve psychiatric diagnosis.更多还原