| 引用本文: | 张乐,刘忠,张建强,等.基于人工蜂群算法优化的改进高斯过程模型.[J].国防科技大学学报,2014,36(1):154-160.[点击复制] |
| ZHANG Le,LIU Zhong,ZHANG Jianqiang,et al.Optimized improved Gaussian process model based on artificial bee colony algorithm[J].Journal of National University of Defense Technology,2014,36(1):154-160[点击复制] |
|
| |
|
|
| 本文已被：浏览 9419次 下载 7495次 |
| 基于人工蜂群算法优化的改进高斯过程模型 |
| 张乐1,2, 刘忠1, 张建强1, 任雄伟1 |
|
(1.海军工程大学 电子工程学院，湖北 武汉 430033;2.
2.武汉军械士官学校，湖北 武汉 430075)
|
| 摘要: |
| 高斯过程(GP)的非线性特征导致其对大样本的训练时间复杂度过高，而且其超参数的选取是否适当直接影响高斯过程回归模型的预测精度。提出采用人工蜂群（ABC）算法优化改进GP以减小时间复杂度和提高预测精度。改进GP通过选取训练样本的子样本进行模型学习，以降低训练过程的时间复杂度。ABC通过优化改进GP的超参数，提升预测精度。选取训练样本的子样本构建改进GP回归（GPR）模型，采用ABC算法搜寻改进GPR的最优超参数，并用得到的超参数构建最优的改进GPR模型，输入测试样本进行预测并输出预测精度。将该模型应用于解决海上远程精确打击（LPSS）体系作战效能评估问题中，通过MATLAB仿真实验，与常见的多种优化方法相比较，验证了该模型的有效性。 |
| 关键词: 改进高斯过程 人工蜂群算法 超参数 参数优化 |
| DOI：10.11887/j.cn.201401027 |
| 投稿日期：2013-08-01 |
| 基金项目:国家部委资助项目（513300303） |
|
| Optimized improved Gaussian process model based on artificial bee colony algorithm |
| ZHANG Le1,2, LIU Zhong1, ZHANG Jianqiang1, REN Xiongwei1 |
|
(1.College of Electronic, Naval University of Engineering, Wuhan 430033, China;2. Wuhan Mechanical Technology College, Wuhan 430075,China)
|
| Abstract: |
| Gaussian Process (GP) is characterized by the non-linear property, which leads to too high training time complexity for a large sample, And the hyper-parameters directly affect the prediction accuracy of Gaussian Process. The method of improved GP optimized by the artificial bee colony (ABC) algorithm is proposed to reduce the time complexity and to improve the prediction accuracy. Improved GP constructs the model by selecting a sub-sample of training samples to reduce training time. ABC optimizes the hyper-parameters of improved GP to improve prediction accuracy. Firstly, the improved GPR model is constructed by selecting a sub-sample of training samples; then it is followed by ABC algorithm searching the optimal hyper-parameters of improved GPR; finally the test sample is used to predict and output the prediction accuracy. The model is applied to solve maritime long-range precision sea strike (LPSS) system-of-systems operational effectiveness evaluation issues, and the MATLAB simulation experiments verify the validity of the model compared with other evolutional algorithms. |
| Keywords: improved gaussian process artificial bee colony algorithm hyper-parameters parameters optimization |
|
|
|
|
|
