基于最大熵方法对测量数据估计的改进方法研究  

Study on Evaluation of Measurement Result Based on Improved Maximum Entropy Method

作　　者：贾波[1] 刘福[1] 雷正伟[1] 

机构地区：[1]军械工程学院,石家庄050003

出　　处：《价值工程》2010年第28期228-229,共2页Value Engineering

摘　　要：针对未知概率分布时难以精确估算测量结果的问题,利用最大熵值法以最少的主观成分拟合测量数据概率分布。在仿真实现过程中,发现最大熵方法存在溢出的问题,通过变量变化法成功解决了此问题。最后,通过MATLAB仿真实现,证明了改进的算法更好的扩展了其适用范围,其估算结果具有可靠性。In the analysis of measurement data with unknown probability distribution,it is difficult to accurately evaluate the measurement result,through using the maximum entropy method to fit the least subjective probability distribution of measurement data.In the simulation process,the author found that the maximum entropy method exists overflow problem,and the problem was successfully resolved by variable change method.Finally,through MATLAB simulation,improved algorithm is proved to better expand its scope of application,and the estimation result is reliable.

关 键 词：最大熵 仿真 MATLAB 

分 类 号：TH71[机械工程—测试计量技术及仪器]