首页> 外国专利> METHOD FOR CODING NUMBERS IN HOP FIELD NEURAL NETWORK

METHOD FOR CODING NUMBERS IN HOP FIELD NEURAL NETWORK

机译:跳神经网络中数字编码的方法

摘要

A kind of purpose: method, for quantity encoding setting to be encoded at by necessary quantity, use the neuron of a neural network, the problem of increasing neural network by an extra neuron-capacity of decomposition, many mapping process represent compiling method for simple bringing for sum, and a group/weight represents compiling method. Construction: a kind of method represents cataloged procedure containing a binary system, a simple sum represents cataloged procedure, and a group/weight represents cataloged procedure for drawing a neuron state space to an integer state space. Binary system represents cataloged procedure and is equal to the binary system converting process for being used for a digital computer. The number of one entire neuron is as described below; Q is equal to increase record 2 and is added to 1 by N. The number of entire neuron be used to show that in binary system, number N represents cataloged procedure. The simple sum that number N is shown as the neuron state in simple sum represents cataloged procedure. Group/weight represents cataloged procedure with a binary system and an a kind of mix type of simple total number system.
机译:一种目的:方法是,将数量编码设置为按必要的数量进行编码,使用神经网络的神经元,通过额外的神经元分解能力增加神经网络的问题,许多映射过程代表了简单的编译方法求和,组/权重代表编译方法。构造:一种方法表示包含二进制系统的分类过程,简单的和表示分类过程,而组/权重表示将神经元状态空间绘制为整数状态空间的分类过程。二进制系统代表分类的过程,并且等同于用于数字计算机的二进制系统转换过程。一个完整的神经元的数量如下所述; Q等于增加记录2,并加N到1。整个神经元的数量用于显示在二进制系统中,N代表分类过程。 N以简单总和表示为神经元状态的简单总和表示分类的过程。组/权重代表具有二进制系统和一种简单总数系统的混合类型的分类程序。

著录项

  • 公开/公告号KR20000038508A

    专利类型

  • 公开/公告日2000-07-05

    原文格式PDF

  • 申请/专利权人 SEO TAE JUN;

    申请/专利号KR19980053524

  • 发明设计人 SEO TAE JUN;

    申请日1998-12-08

  • 分类号G06F17/00;

  • 国家 KR

  • 入库时间 2022-08-22 01:45:41

相似文献

  • 专利
  • 外文文献
  • 中文文献
获取专利

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号