| 神经网络集成的免疫学习算法 |
| |
| 引用本文： | 孟江,王耀才,王天成,巩敦卫.神经网络集成的免疫学习算法[J].中国矿业大学学报,2005,34(4):486-489. |
| |
| 作者姓名： | 孟江 王耀才 王天成 巩敦卫 |
| |
| 作者单位： | 中国矿业大学,信息与电气工程学院,江苏,徐州,221008 |
| |
| 基金项目： | 国家自然科学基金项目(60304016) |
| |
| 摘 要： | 针对神经网络集成中个体独立训练的低效性，提出一种神经网络集成的免疫学习算法NEIL，通过对集成单体神经网络的免疫优化，借助免疫算法的多峰值搜索和并行优化特性，将最终的收敛抗体群分别对应神经网络集成的各个单体，实现集成的同时训练过程，仿真结果表明NEIL算法是有效的，既保持了启发式优化方法的并行性，提高了神经网络集成的学习效率，又可保证单体网络之间相互独立，具有较大的差异度，进一步增强神经网络集成的泛化性能。
|
| 关 键 词： | 神经网络集成 同时训练 免疫算法 泛化 差异度 |
| 文章编号： | 1000-1964(2005)04-0486-04 |
| 修稿时间： | 2004年10月2日 |
Immune Learning Algorithm Based on Neural Network Ensemble |
| |
| MENG Jiang,WANG Yao-cai,WANG Tian-Cheng,GONG Dun-wei.Immune Learning Algorithm Based on Neural Network Ensemble[J].Journal of China University of Mining & Technology,2005,34(4):486-489. |
| |
| Authors: | MENG Jiang WANG Yao-cai WANG Tian-Cheng GONG Dun-wei |
| |
| Abstract: | Aimed at low efficiency of individual training in neural network ensemble (NNE), a neural-network ensemble immune learning (NEIL) algorithm is presented, which adopts immune algorithm (IA) to optimize individuals in NNE, treats convergent antibodies of IA as the corresponding neural network, and realizes the simultaneous training of NNE, according to the characteristics of multi-peaked search and parallel optimization of IA. The simulating result shows that the NEIL algorithm is of availability, which not only increases the learning efficiency of NNE based on the parallelism of heuristic optimization methods, but also improves the NNE generalization with bigger ambiguities because of the greater independence among NNE individuals. |
| |
| Keywords: | neural network ensemble simultaneous training immune algorithm generalization ambiguity |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
