摘要
将计算机发生的“扰动-响应”曲线的基值与正态分布伪随机数迭加,可以获得人工模拟的“扰动-响应”实验曲线。本文分别运用矩量法、加权矩量法、传递函数法、傅立叶分析法及时间域最小二乘拟合法对模拟实验曲线加以处理与参数估值,并采用 Monte-Carlo 法对这五种数据处理方法作出比较与评价。模拟实验计算结果表明,傅立叶分析法和时间域最小二乘拟合法精度最高,是值得推荐的数据处理方法。
By summing the calculated value of“stimulus-response”curve and the normalpseudo-random number,the computer-simulated“stimulus-response”experimentalcurves were obtained.Then these simulated curves were treated with five methods:mo-ment,weighted moment,transfer function,Fourier analysis and least squares fitting inthe time domain for parameter estimation.Monte-Carlo method was employed in evalu-ating these five methods.The results showed that Fourier analysis method and least squaresfitting in the time domain method were superior to the others and could be recommended.
出处
《计算机与应用化学》
CAS
CSCD
1989年第4期254-259,共6页
Computers and Applied Chemistry
