The least trimmed squares estimator (LTS) is a well known robust estimator in terms of protecting the estimate from the outliers. Its high computational complexity is however a problem in practice. We show that the LT...The least trimmed squares estimator (LTS) is a well known robust estimator in terms of protecting the estimate from the outliers. Its high computational complexity is however a problem in practice. We show that the LTS estimate can be obtained by a simple algorithm with the complexity 0( N In N) for large N, where N is the number of measurements. We also show that though the LTS is robust in terms of the outliers, it is sensitive to the inliers. The concept of the inliers is introduced. Moreover, the Generalized Least Trimmed Squares estimator (GLTS) together with its solution are presented that reduces the effect of both the outliers and the inliers. Keywords Least squares - Least trimmed squares - Outliers - System identification - Parameter estimation - Robust parameter estimation This work was supported in part by NSF ECS — 9710297 and ECS — 0098181.展开更多
Nonlinear least trimmed squares (NLTS) estimator is a very important kind of nonlinear robust estimator, which is widely used for recovering an ideal high-quality signal from contaminated data. However, the NLTS est...Nonlinear least trimmed squares (NLTS) estimator is a very important kind of nonlinear robust estimator, which is widely used for recovering an ideal high-quality signal from contaminated data. However, the NLTS estimator has not been widely used because it is hard to compute, This paper develops an algorithm to compute the NLTS estimator based on a random differential evolution (DE) strategy. The strategy which uses random DE schemes and control variables improves the DE performance. The simulation results demonstrate that the algorithm gives better performance and is more convenient than existing computing algorithms for the NLTS estimator. The algorithm makes the NLTS estimator easy to apply in practice, even for large data sets, e.g. in a data mining context.展开更多
文摘The least trimmed squares estimator (LTS) is a well known robust estimator in terms of protecting the estimate from the outliers. Its high computational complexity is however a problem in practice. We show that the LTS estimate can be obtained by a simple algorithm with the complexity 0( N In N) for large N, where N is the number of measurements. We also show that though the LTS is robust in terms of the outliers, it is sensitive to the inliers. The concept of the inliers is introduced. Moreover, the Generalized Least Trimmed Squares estimator (GLTS) together with its solution are presented that reduces the effect of both the outliers and the inliers. Keywords Least squares - Least trimmed squares - Outliers - System identification - Parameter estimation - Robust parameter estimation This work was supported in part by NSF ECS — 9710297 and ECS — 0098181.
基金the Key Technologies Research and Development Program of the Tenth Five-Year Plan of China (No. 2001609A12)
文摘Nonlinear least trimmed squares (NLTS) estimator is a very important kind of nonlinear robust estimator, which is widely used for recovering an ideal high-quality signal from contaminated data. However, the NLTS estimator has not been widely used because it is hard to compute, This paper develops an algorithm to compute the NLTS estimator based on a random differential evolution (DE) strategy. The strategy which uses random DE schemes and control variables improves the DE performance. The simulation results demonstrate that the algorithm gives better performance and is more convenient than existing computing algorithms for the NLTS estimator. The algorithm makes the NLTS estimator easy to apply in practice, even for large data sets, e.g. in a data mining context.