摘要
目的探索高阶神经网络模型中的学习算法。方法通过在高阶神经网络模型中引入新的学习算法增强高阶神经网络模型性能,给出了新学习算法的理论分析、收敛性证明,进行了仿真实验。结果新学习算法下的高阶神经网络模型的性能指标优于传统Hebb型学习算法下的高阶神经网络模型。结论新学习算法下的高阶神经网络模型具有更好的优良特性,有利于其在信息存储、模式识别领域中的应用。
Objective To analyze the learning algorithm in higher- order neural networks model. Methods The performance of higher- order neural networks model was improved by adopting a new learning algorithm. The theoretical analysis of the algorithm was given. The convergence of the algorithm was proved. The numerical experiments were carried out. Results The properties of higher- order neural networks model using new learning algorithm were superior to that of the model using traditional Hebb＇s algorithm. Conclusion The remarkable features enable the higher order neural networks model using new learning algorithm to have good prospects in information storage and pattern recognition.
出处
《徐州医学院学报》
CAS
2006年第4期311-314,共4页
Acta Academiae Medicinae Xuzhou
基金
江苏省高校自然科学研究指导性计划项目(03KJD310224)
关键词
高阶神经网络
模型
学习算法
higher- order neural networks
model
learning algorithm
