| 基于局部平稳的随机序列变化点检测及参数估计 |
| |
| 引用本文： | 王聪,孙晓颖.基于局部平稳的随机序列变化点检测及参数估计[J].吉林大学学报(工学版),2012(Z1):396-400. |
| |
| 作者姓名： | 王聪 孙晓颖 |
| |
| 作者单位： | 清华大学电子工程系;吉林大学通信工程学院 |
| |
| 基金项目： | 国家自然科学基金项目(60940011) |
| |
| 摘 要： | 在Kullback-Leibler对称散度的框架下,建立了检测结构变化点的辨识信息量和确定AR模型阶数的KICC准则,提出了分段拟合AR模型的系统方法,采用混合粒子群优化方法确定变化点的分布及位置,应用最小二乘法估计模型参数,提高了运算速度和精度。仿真实验表明,本文提出的随机序列变化点检测和AR模型参数估计方法,具有高度的稳定性和可靠性,是一种有效的方法。
|
| 关 键 词： | 信息处理技术 Kullback-Leibler对称散度 变化点 参数估计 |
Locally stationary based detection of change points in random sequence and estimation of parameters |
| |
| WANG Cong,SUN Xiao-ying.Locally stationary based detection of change points in random sequence and estimation of parameters[J].Journal of Jilin University:Eng and Technol Ed,2012(Z1):396-400. |
| |
| Authors: | WANG Cong SUN Xiao-ying |
| |
| Affiliation: | 1.Department of Electronic Engineering,Tsinghua University,Beijing 100084,China;2.College of Communication Engineering,Jilin University,Changchun 130022,China) |
| |
| Abstract: | Kullback-Leibler symmetric divergence was applied to derive discrimination information of detection of change points and KICC criterion of determination AR model order.Method of fitting piecewise AR models was proposed,hybrid particle swarm optimization was used to determine distribution and location of change points,the least squares was applied to estimate model parameters,computing speed and accuracy were improved.Simulation showed that the method has higher stability and reliability,and is an effective method. |
| |
| Keywords: | information technology Kullback-Leibler symmetric divergence change point parameter estimation |
| 本文献已被 CNKI 等数据库收录！ |
|
