CAJ | 학술논문

将传统模式识别神经网络与人类认知的方式相结合,提出了结构开放式模式识别神经网络的概念,并设计了结构开放式模式识别神经网络的体系结构.与传统模式识别神经网络相比,这种神经网络不但提高了神经网络的学习记忆功能,而且在利用计算机实现智能化时大大减少了数据结构的数据量.为了能够更容易的理解这种网络,并针对一类特殊情况的应用进行研究,设计了一个结构开放式模式识别神经网络的结构和操作流程,并通仿真试验,证明了这种网络的有效性和实用性.
장전통모식식별신경망락여인류인지적방식상결합,제출료결구개방식모식식별신경망락적개념,병설계료결구개방식모식식별신경망락적체계결구.여전통모식식별신경망락상비,저충신경망락불단제고료신경망락적학습기억공능,이차재이용계산궤실현지능화시대대감소료수거결구적수거량.위료능구경용역적리해저충망락,병침대일류특수정황적응용진행연구,설계료일개결구개방식모식식별신경망락적결구화조작류정,병통방진시험,증명료저충망락적유효성화실용성.
This paper presents the conception of the Open Structure Mode Recognition Neural Network (OSMRNN) based the general mode recognition neural network and the cognize of the human. And the system structure of the OSMRNN is designed in this paper. The OSMRNN not only improves the learning and memory ability, but also reduces the data number on computer to general mode recognition neural network. In order to comprehend the network easily, this paper designed a structure of the OSMRNN and the process of the operation for one special inputting. And the OSMRNN is confirmed efficient by one example.