| 基于SVM的概率密度估计及分布估计算法 |
| |
| 引用本文： | 徐玉兵,谭瑛,曾建潮,张建华.基于SVM的概率密度估计及分布估计算法[J].计算机与数字工程,2009,37(6):25-28. |
| |
| 作者姓名： | 徐玉兵 谭瑛 曾建潮 张建华 |
| |
| 作者单位： | 太原科技大学系统仿真与计算机应用研究所,太原,030024 |
| |
| 摘 要： | 在最大熵分布估计算法中，根据Jaynes原理来建立分布估计算法中的概率密度。基于SVM的概率密度估计则是根据概率密度的定义，由核函数构造一个包含未知参数的概率密度函数。它根据样本点建立这个概率密度的数学规划模型，并用不敏感损失函数的支持向量机方法来求解这个模型。对得到的概率密度进行仿真测试，最后将得到的密度应用到分布估计算法中。
|
| 关 键 词： | 核函数 样本点 舍选法 分布估计算法 |
Density Estimation Based on Support Vector Machine with its Application in Estimation of Distribution Algorithms |
| |
| Xu Yubing,Tan Ying,Zeng Jianchao,Zhang Jianhua.Density Estimation Based on Support Vector Machine with its Application in Estimation of Distribution Algorithms[J].Computer and Digital Engineering,2009,37(6):25-28. |
| |
| Authors: | Xu Yubing Tan Ying Zeng Jianchao Zhang Jianhua |
| |
| Affiliation: | Systems Simulation and Computer Application Institute;Taiyuan University of Science and Tenchnology;Taiyuan 030024 |
| |
| Abstract: | In the maximum entropy distribution estimate algorithm and its application,it constructs the density of the estimation of distribution algorithms according to Jaynes principle.On the basis of probability density and through the kernel function,density estimation based on support vector machine constructs a probability density function which includes unknown parameters.It takes use of insensitive loss function's SVM to solve the mathematical model that was constructed by sample points.Having a simulation tes... |
| |
| Keywords: | kernel function empirical distribution function regression estimate acceptance-rejection method estimation of distribution algorithms |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
