林励莉, 王涛. 推进学习对传神经网络ACPN[J]. 华南师范大学学报(自然科学版), 2011, (2).
引用本文: 林励莉, 王涛. 推进学习对传神经网络ACPN[J]. 华南师范大学学报(自然科学版), 2011, (2).
Improved counterpropagation networks with adaptive learning strategy[J]. Journal of South China Normal University (Natural Science Edition), 2011, (2).
Citation: Improved counterpropagation networks with adaptive learning strategy[J]. Journal of South China Normal University (Natural Science Edition), 2011, (2).

推进学习对传神经网络ACPN

Improved counterpropagation networks with adaptive learning strategy

  • 摘要: 结合Adaboost算法的加权投票机制,提高对传神经网络算法CPN(Counterpropagation Networks)的学习效率,提出新型快速分类算法ACPN.实验证明,新算法的学习最小误差比传统CPN算法下降了96%,训练时间同比下降44%,网络训练阶段误差下降趋势明显稳定,不像传统CPN有一定的波动性.

     

    Abstract: This study presents a novel Adaptive boosting theory-Counterpropagetion neural network(ACPN) for solving forecasting problems. The boosting concept is integrated into the CPN learning algorithm for learning effectively. Compared with traditional CPN, the minimum training error and learning time in ACPN network fell about 96% and 44%. Furthermore, the curve of trainning error in ACPN presents downtrend basically and has less fluctuation.

     

/

返回文章
返回