LVQ神经网络在临床辅助鉴别诊断中的应用研究 - 中国数字医学.PDFVIP

数字医学研究与应用 Research snd Application of Digital Medicine LVQ神经网络在 临床辅助鉴别诊断中的应用研究 李文念① 孙卫② 李颖① 俞思伟③ 摘 要 选取协和医院64例结节病和42例肺结核病病历作为数据样本，探究建立基于LVQ神经网络的临床辅助鉴别诊断模 型以实现这两种疾病的鉴别诊断，测试结果总体确诊率达到87.50%，并与BP神经网络算法结果相对比。实证分析结果表 明，该模型对疾病的辅助鉴别诊断有着良好的应用效果，利用LVQ神经网络实现疾病的临床辅助鉴别诊断是有效的。 关键词 LVQ神经网络 鉴别诊断 结节病 肺结核 Doi:10.3969/j.issn.1673-7571.2013.05.020 Application of Clinical Auxiliary Differential Diagnosis Based on Learning Vector Quantization Neural Network / LI Wen-nian, SUN Wei, LI Ying, et al//China Digital Medicine.-2013 8(5): 72 to 75 Abstract This paper put forward a clinical auxiliary differential diagnosis model based on learning vector quantization(LVQ) neural network．An experiment was designed with the data sample of 106 medical records consisting of 64 patients with sarcoidosis and 42 patients with tuberculosis from the Union Hospital. The empirical analysis results show that the test results have an overall positive rate of 87.50%, and the model for the clinical auxiliary differential diagnosis of diseases has a good application effect. The clinical auxiliary differential diagnosis based on LVQ neural network is appropriate, and it can make differential diagnosis of diseases effectively. Keywords LVQ neural network, differential diagnosis, Sarcoidosis, Tuberculosis Fund project National National Key Technology RD Program for the 12th five-year plan-Study on key technologies of large-scale semantics computation for foreign language scientific and technological knowledge organization system (No. 2011BAH10B04-04) Corresponding author Institute of Scientific and Technical Information of China, Beijing 100038, P.R.C. 上世纪50年代末，F·Rosenblatt设计并制作了多层感知神经网络，首次把人工神经网络的研究从理论探讨付诸工程实 践。从此，人工神经网络的发展跌宕起伏，至今已有很多神经网络模型，深入应用在模式识别、知识工程、决策支持等各