CAJ | 학술논문

为验证单RBF神经网络更适用于老年痴呆症的预测诊断，通过仿真实验将单BP神经网络、单RBF神经网络、遗传算法优化BP神经网络及遗传算法优化RBF神经网络分别应用于老年痴呆症的预测诊断，建立这四种网络模型，并对四种网络模型的预测结果进行分析比较。仿真实验在Matlab软件平台上进行。结果表明：在老年痴呆症的预测诊断中，单RBF神经网络比单BP神经网络预测结果更好，建模时间更短。此外，单RBF神经网络与遗传算法优化的BP神经网络预测结果相同，但单RBF神经网络建模较为简单，预测结果更为稳定。而遗传算法对RBF神经网络优化作用不明显。因此，单RBF神经网络更适用于老年痴呆症的预测诊断，实际应用时可以此结论作为理论指导。
위험증단RBF신경망락경괄용우노년치태증적예측진단，통과방진실험장단BP신경망락、단RBF신경망락、유전산법우화BP신경망락급유전산법우화RBF신경망락분별응용우노년치태증적예측진단，건립저사충망락모형，병대사충망락모형적예측결과진행분석비교。방진실험재Matlab연건평태상진행。결과표명：재노년치태증적예측진단중，단RBF신경망락비단BP신경망락예측결과경호，건모시간경단。차외，단RBF신경망락여유전산법우화적BP신경망락예측결과상동，단단RBF신경망락건모교위간단，예측결과경위은정。이유전산법대RBF신경망락우화작용불명현。인차，단RBF신경망락경괄용우노년치태증적예측진단，실제응용시가이차결론작위이론지도。
In order to verify single RBF neural network is more suitable for the predictive diagnosis of senile dementia, through the simulation experiment, a single BP neural network, a single RBF neural network, a genetic algorithm to optimize BP neural network and a genetic algorithm to optimize RBF neural network are used to predict senile dementia, establishing of these four kinds of network model, then analyzing and comparing the forecasted results of these four kinds of network model. The simulation experiments were carried out on the platform of Matlab software, the results show that: in the predictive diagnosis of senile dementia, the single RBF neural network predictive results is higher than the single BP neural network,and the modeling time is shorter. Furthermore, the prediction results of the single RBF neural network is as the same as the genetic algorithm to optimize BP neural network, but the single RBF neural network model is relatively simple, and the prediction results are more stable. Therefore, diagnosis and prediction of the single RBF neural network is more suitable for senile dementia, and this conclusion can be used as a theoretical guide to the actual application.