Two Iterative Reweighted Algorithms for Systems Contaminated by Outliers
- 24 August 2023
- journal article
- research article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 72 (00189456) , 1-11
Abstract
This study proposes two iterative reweighted algorithms for systems whose data are contaminated by outliers. For the negative effect caused by the outliers, traditional least squares and gradient descent algorithms cannot obtain unbiased estimates, while the variational Bayesian and expectation-maximization algorithms have the assumption that the prior knowledge of the outlier is available. To deal with these dilemmas, two iterative reweighted algorithms are developed. By assigning suitable weights for each data set, unbiased parameter estimates can be obtained. In addition, the weights of the corrupted data sets become smaller and smaller with the increased number of iterations, and then the contaminated data can be picked out from the data sets. The proposed algorithms do not require the prior knowledge of the outliers. Convergence analysis and numerical experiments show effectiveness of the iterative reweighted algorithms.Keywords
Funding Information
- National Natural Science Foundation of China (61973137)
- Natural Science Foundation of Jiangsu Province (BK20201339)
This publication has 31 references indexed in Scilit:
- Expectation maximization estimation algorithm for Hammerstein models with non-Gaussian noise and random time delay from dual-rate sampled-dataDigital Signal Processing, 2018
- Iterative Identification of Hammerstein Parameter Varying Systems With Parameter Uncertainties Based on the Variational Bayesian ApproachIEEE Transactions on Systems, Man, and Cybernetics: Systems, 2017
- Subspace Identification of Individual Systems Operating in a Network (SI $^2$ON)IEEE Transactions on Automatic Control, 2017
- Robust Estimation of ARX Models With Time Varying Time Delays Using Variational Bayesian ApproachIEEE Transactions on Cybernetics, 2017
- Decision-Based System Identification and Adaptive Resource AllocationIEEE Transactions on Automatic Control, 2016
- Iterative Maximum Likelihood and Outlier-robust Bipercentile Estimation of Parameters of Compound-Gaussian Clutter With Inverse Gaussian TextureIEEE Signal Processing Letters, 2016
- Robust multiple-model LPV approach to nonlinear process identification using mixture t distributionsJournal of Process Control, 2014
- System Identification Via Sparse Multiple Kernel-Based Regularization Using Sequential Convex Optimization TechniquesIEEE Transactions on Automatic Control, 2014
- Partially Coupled Stochastic Gradient Identification Methods for Non-Uniformly Sampled SystemsIEEE Transactions on Automatic Control, 2010
- Iteratively reweighted least squares minimization for sparse recoveryCommunications on Pure and Applied Mathematics, 2009