基于RBF的肾阳虚大鼠甲状腺与性腺功能的动态分析

Dynamic analysis of thyroid gland and gonad function on the rat of kidney-yang deficiency based on radial basis function neural networks

目的通过建立肾阳虚大鼠三碘甲腺原氨酸(T3)、甲状腺素(T4)与睾酮(T)的人工神经网络模型,探讨肾阳虚过程中甲状腺与性腺功能的动态关系。方法肌肉注射氢化可的松复制肾阳虚大鼠模型,分别观察早期、中期和晚期肾阳虚大鼠T3、T4与T等指标,建立T3、T4与T之间的径向基函数网络模型,利用模型分析甲状腺与性腺功能之间的变化规律和特征。结果肾阳虚网络模型的相对均方误差都在10%以内。结论径向基函数网络模型能比较准确反映肾阳虚大鼠甲状腺与性腺功能之间复杂的非线性关系。

Objective To study the dynamic correlations between thyroid and gonad functions during kidney-yang deficiency with radial basis function neural network. Methods The rat model of kidney -yang deficiency was duplicated by intramuscular injection of cortisone. Indices of thyroid and gonad function on the rats of kidney-yang deficiency were observed respectively at early, middle and late periods respectively. The experimental data was conformed to radial basis function neural networks （ RBF ） , and the rule and characteristic of changes between thyroid and gonad function were analyzed with RBF models. Results The mean square errors on RBF models of kidney-yang deficiency were below 10 percent. Conclusion The RBF models show the excellent ability of self-organization and self-adaption, and it conforms to the complicated non-linear relation between thyroid and gonad function during kidney-yang deficiency.