Elman神经网络改进模型在脑膜炎诊断中的应用

Application of improved Elman neural network model in diagnosis of meningitis

为了客观地反映脑脊液(Cerebrospinal Fluid,CSF)所蕴涵的病理信息,研究CSF多变量非线性诊断指标和脑膜炎分类标准的内在联系,提出一种Elman神经网络改进模型辅助诊断脑膜炎疾病的方法。构建2层Elman神经网络改进模型网络训练和仿真的实验结构,分别把85和51例临床病例确诊数据作为训练样本和仿真数据的输入。仿真结果显示,采用Elman神经网络的改进模型应用于脑膜炎疾病的辅助诊断可以达到均方误差10-2精度。Elman神经网络改进模型针对CSF复杂数据关系辅助诊断脑膜炎疾病的智能计算是可行的。

An improved Elman neural network model for auxiliary diagnosis of meningitis disease is put forward, which aims to show the pathological information of the cerebrospinal fluid and the inner relationship between the index of cere-brospinal fluid nonlinear multivariable diagnosis and the criteria of meningitis classification. A two-level training and sim-ulation experimental structure of improved Elman neural network model network is constructed, and 85 and 51 clinical cases confirmed data as training samples and simulation data are inputted respectively. The simulation results show that, with the improved Elman neural network model, the auxiliary diagnosis mean squared error accuracy can reach 10-2. The intelligent computation of improved Elman neural network model in aided diagnosis of meningococcal disease is feasible.