| 基于粒子群算法的神经网络短期降水预报建模研究 |
| |
| 引用本文： | 吴建生. 基于粒子群算法的神经网络短期降水预报建模研究[J]. 智能系统学报, 2006, 1(2): 67-73 |
| |
| 作者姓名： | 吴建生 |
| |
| 作者单位： | 柳州师范高等专科学校,数学与计算机科学系,广西,柳州,545004 |
| |
| 基金项目： | 广西省教育厅资助项目（200508234）. |
| |
| 摘 要： | 用多样性粒子群算法优化神经网络的网络结构和连接权，获得神经网络集成个体；进一步用二次规划方法，计算各集成个体的最优非负权系数进行组合集成，生成神经网络集成的输出结论，进行短期降水预报建模研究．以广西全区的月降水量实例分析，结果表明该方法能有效提高系统的泛化能力．
|
| 关 键 词： | 神经网络集成 粒子群优化 二次规划 |
| 文章编号： | 1673-4785（2006）02-0067-07 |
| 收稿时间： | 2006-04-28 |
| 修稿时间： | 2006-04-28 |
Study on the short-time rainfall prediction model of neural ensemble based on PSO algorithms |
| |
| WU Jian-sheng. Study on the short-time rainfall prediction model of neural ensemble based on PSO algorithms[J]. CAAL Transactions on Intelligent Systems, 2006, 1(2): 67-73 |
| |
| Authors: | WU Jian-sheng |
| |
| Affiliation: | Department of Mathematics and Computer Science, Liuzhou Teachers College, Liuzhou 545004,China |
| |
| Abstract: | This paper versity-guided Partic presents the evolving neural network architecture and connection weights based on Dile Swarm Optimization algorithms. The ensemble strategy is carried out by using the quadratic programming to calculate the best non-negative weights. The weighted coefficient of each ensemble individual is obtained. This method can be used to establish the forecast model of the short-time rainfall. The applied example is built with the monthly mean rainfall in the whole area of Guangxi. The result shows that this method can effectively increase the generalization ability of neural network. |
| |
| Keywords: | neural network ensemble particle swarm optimization quadratic program |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
