CAJ | 학술논문

神经网络是目前处理科技领域各类问题的一个重要工具，它由大量简单单元以及这些单元的分层组织大规模联结而成，力图像生物神经系统一样处理事物； BP网络采用传播算法，是目前应用最为广泛和可靠的神经网络之一，具有较强的分类和学习能力。从模式识别出发，在选取典型实例的基础上，建立BP网络算法模型，对算法进行动态误差修正的改进，提高了算法的收敛速度，并根据算法流程，通过运用Matlab软件对其进行仿真验证，说明了BP网络算法在模式识别中具有应用可行性。
신경망락시목전처이과기영역각류문제적일개중요공구，타유대량간단단원이급저사단원적분층조직대규모련결이성，력도상생물신경계통일양처리사물； BP망락채용전파산법，시목전응용최위엄범화가고적신경망락지일，구유교강적분류화학습능력。종모식식별출발，재선취전형실례적기출상，건립BP망락산법모형，대산법진행동태오차수정적개진，제고료산법적수렴속도，병근거산법류정，통과운용Matlab연건대기진행방진험증，설명료BP망락산법재모식식별중구유응용가행성。
Neural network is an important tool for processing technology fields of various kinds of issues, which is composed by a large number of simple units and the hierarchical organization of mass coupling of these units. Force diagram biological deal with things like neural system; BP network use propagation algorithm, which is one of the most widely used and reli-able neural networks with strong classification and learning ability. Based on pattern recognition and selection of typical ex-amples, the paper was established BP network algorithm model, improved dynamic error correction of algorithm, which pro-moted rate of convergence of algorithm. According to the algorithm process, the paper was explained feasibility of applica-tions of BP network algorithm in pattern recognition through using the Matlab software simulation.