摘要: 讨论了将粒子信息存入神经网络的机制,以加速网络训练,提高清晰度,提出了语言权重的GNN及语言算法规则。GNN比标准BP网络需要少训练时间来学习已知问题,同时知识的表示更易于理解。
关键词:
粒子神经网络,
粒子计算,
语言变量
Abstract: This article discusses the mechanism which graular information will be stored in neural network,in order to speed up network training and increase clarity,then presents GNN based on linguistic weights and linguistic algorithm rules.
中图分类号:
丁雪梅,伦立军,李英梅. 神经网络中的粒子权重计算[J]. , 2005, 31(04): 177-179.
DING Xuemei,LUN Lijun,LI Yingmei. Granular Weights Computing in Neural Networks[J]. , 2005, 31(04): 177-179.