【摘要】 为使原始LTS(least trimm ed squares)方法能够处理非线性问题,研究非线性LTS稳健估计方法。说明该方法的解一定是部分观测值的非线性最小二乘估计。该方法可通过求解非线性最小二乘问题得到确切解。基于MM EA(m in im um m ax im um exchange a lgorithm)算法和非线性最小二乘技术,构建求解非线性LTS估计近似解的算法。仿真结果表明非线性LTS估计方法能够同时抵抗来自X方向和Y方向的多个异常,与传统方法相比具有更好的稳健性。更多还原

【Abstract】 A robust nonlinear LTS(least trimmed squares) estimation method was developed for nonlinear model parameter estimation problems.The method seeks the solution of the nonlinear least squares estimate of partially observed values.The algorithm to calculate the approximate nonlinear LTS solution uses the minimum maximum exchange algorithm and the nonlinear least squares technique.Simulation results indicate that the nonlinear LTS is able to reject outliers in both the X and Y directions so it is more robust than the classic method.更多还原