应用RBF激励WASD神经网络估算GFR

价格低廉与高准确率的矛盾是测量肾小球滤过率（glomerular filtration rate，GFR）中遇到的主要难题。采用径向基函数（radial basis function，RBF）神经网络和权值与结构确定法（weights-and-structure-determination，WASD）相结合的方法，并基于中山大学附属第三医院的患者数据进行神经网络建模，对肾病患者进行肾小球滤过率估算。计算机数值实验结果显示该方法在50%符合率标准下能达到90%的准确率，而传统方程中最优的准确率为68%。

Application of RBF-Activated WASD Neuronet in Estimating GFR

The contradiction between low price and high accuracy is the main problem encountered in estimating glomerular filtration rate (GFR). With the combination of RBF neuronet and weights-and-structure-determination (WASD) algorithm, and based on the data of patients from the Third Affiliated Hospital of Sun Yat-sen University, a neuronet model was built to estimate chronic kidney disease (CKD) patients'' GFR. Numerical-experiment results show that the presented method can reach 90% accuracy with 50% accuracy as a standard, while the highest accuracy of the traditional equations is 68%.