摘要
函数连接型神经网络是一种新型的单层网络。应用函数连接型网络成功地实现了根据人体头发中微量元素的含量对老年痴呆症的诊断，表明该种网格性能良好，预测精度高。它为疾病的诊断提供了一种新的途径。
The FunctionalLink net (FLN),a new type of singlelayer network,is introduced and used to diagnose the senile dementia by the real data about the trace element contents in human hair.The original input patterns are enhanced nonlinearly as the new inpur of FLN,and this decreases the probability of the net in falling to local minima.With a small amount of training data,the experimental results show that FLN possesses a faster learning speed and a good generalization ability.This suggests that FLN be an effective tool in the diagnosis of diseases in the future.
出处
《微量元素与健康研究》
CAS
1999年第2期21-23,共3页
Studies of Trace Elements and Health
基金
国家自然科学基金
