Performance analysis of the adaptive algorithm for bias-to-variance tradeoff

• Published in: IEEE transactions on signal processing Vol. 52; no. 5; pp. 1228 - 1234
• Main Author: Stankovic, L.
• Format: Journal Article
• Language: English
• Published: New York, NY IEEE 01.05.2004; Institute of Electrical and Electronics Engineers; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Adaptive algorithm; Adaptive algorithms; Adaptive signal processing; Algorithm design and analysis; Algorithms; Applied sciences; Confidence intervals; Detection, estimation, filtering, equalization, prediction; Exact sciences and technology; Filtering; Filtration; Frequency estimation; Gaussian; Information, signal and communications theory; Multidimensional signal processing; Noise; Optimization; Performance analysis; Signal and communications theory; Signal processing algorithms; Signal representation. Spectral analysis; Signal, noise; Smoothing methods; Telecommunications and information theory; Time frequency analysis; Performance evaluation; Time-frequency analysis; statistical signal processing; Parameter estimation; Filtering; Adaptive algorithm; Adaptive algorithms; Signal estimation; Intercarrier interference; spectral analysis; Optimal estimation; Laplacian; Optimization; Mean square error; Confidence interval; estimation; Gaussian noise; Algorithm performance; Spectrum analysis; Signal processing; noise; Signal analysis; Algorithm analysis
• ISSN: 1053-587X; 1941-0476
• DOI: 10.1109/TSP.2004.826179

An algorithm for the mean squared error (MSE) minimization, through the bias-to-variance ratio optimization, has been recently proposed and used in the literature. This algorithm is based on the analysis of the intersection of confidence intervals (ICIs). The algorithm does not require explicit knowledge of the estimation bias for a "near to optimal" parameter estimation. This paper presents a detailed analysis of the algorithm performances, including procedures and relations that can be used for a fine adjustment of the algorithm parameters. Reliability of the algorithm is studied for various kinds of estimation noise. Results are confirmed on a simulated example with uniform, Gaussian, and Laplacian noise. An illustration of the algorithm application on a simple filtering example is given.