| 基于交叉验证的BP算法的改进与实现 |
| |
| 引用本文： | 韩萌,丁剑. 基于交叉验证的BP算法的改进与实现[J]. 计算机工程与设计, 2008, 29(14) |
| |
| 作者姓名： | 韩萌 丁剑 |
| |
| 作者单位： | 西北第二民族学院计算机科学与技术系,宁夏银川750021;北京交通大学计算机与信息技术学院,北京100044;西北第二民族学院计算机科学与技术系,宁夏银川,750021 |
| |
| 摘 要： | 针对BP算法存在的收敛速度慢等问题提出改进方案,修改其相关参数并且提出如何选择合适的隐藏层节点个数.同时针对学习样本数据的有限性、BP算法易陷入局部最小值和容易出现过拟合等问题进行了研究,提出了采用多重交叉验证的再改进BP算法.仿真结果表明,交叉验证BP算法提高了网络学习的效率.
|
| 关 键 词： | 神经网络 BP算法 交叉验证 过拟合 隐藏层 误差函数 |
Improvement of BP algorithm based on cross-validation method and its implementation |
| |
| HAN Meng,DING Jian. Improvement of BP algorithm based on cross-validation method and its implementation[J]. Computer Engineering and Design, 2008, 29(14) |
| |
| Authors: | HAN Meng DING Jian |
| |
| Affiliation: | HAN Meng1,2,DING Jian2(1.Department of Computer Science , Technology,Second Northwest University for Nationality,Yinchuan 750021,China,2.College of Computer , Information Technology,Beijing Jiaotong University,Beijing 100044,China) |
| |
| Abstract: | BP algorithm exists against the slow convergence and other problem,in order to improve the program,the system amends its relevant parameters and presents the method to choose a suitable number of hidden layer nodes.Focus on learning of the limited nature of the data sample and BP algorithm easy to fall into the local minimum and prone to over-fit problems,the system presented by multiple cross-validation method.The simulation results show that cross-validation BP algorithm improved the efficiency of learnin... |
| |
| Keywords: | neural network BP algorithm cross-validation over-fitting hidden layer error function |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
