论文部分内容阅读
本文以天然留兰香的组分构成与其品质的关系为例,讨论人工神经元网络方法用于复杂信息模式分类的问题.提出一种广义的误差反传训练策略,将网络的训练范围从联接权扩大到神经元模型.这种新的训练方法(GBP)能提高多层前传网络的学习效率,加快收敛的速率.实际运行的结果表明,所需训练时间仅为普通误差反传(BP)训练方法的1/15.并能达到较高的预报精度.
In this paper, the composition of natural Spearmint and its quality as an example, the discussion of artificial neural network method for the classification of complex information patterns.A generalized error backtracking training strategy, the network training range from the connection The new training method (GBP) can improve the learning efficiency and accelerate the convergence rate of multi-layer pre-sent network.The results of actual operation show that the required training time is only BP (Normal Error Back Propagation) Training method of 1 / 15. And can achieve a higher forecast accuracy.