an entropy-like estimator for robust parameter identification一个健壮的参数识别类熵估计量.pdfVIP

Entropy 2009, 11, 560-585; doi:10.3390/ OPEN ACCESS entropy ISSN 1099-4300 /journal/entropy Article An Entropy-Like Estimator for Robust Parameter Identification Giovanni Indiveri Dipartimento Ingegneria Innovazione, University of Salento, Via Monteroni s.n., 73100 Lecce, Italy; E-Mail: giovanni.indiveri@unisalento.it; Tel.: +39 0832 29 7220; Fax +39 0832 29 7733 Received: 7 September 2009 / Accepted: 23 September 2009 / Published: 12 October 2009 Abstract: This paper describes the basic ideas behind a novel prediction error parameter identification algorithm exhibiting high robustness with respect to outlying data. Given the low sensitivity to outliers, these can be more easily identified by analysing the residuals of the fit. The devised cost function is inspired by the definition of entropy, although the method in itself does not exploit the stochastic meaning of entropy in its usual sense. After describing the most common alternative approaches for robust identification, the novel method is presented together with numerical examples for validation. Keywords: system identification; model fitting; data processing Classification: PACS 89.70.-a, 89.70.Cf, 87.19.lo, 07.05.Kf Classification: MSC 93-XX, 93Exx, 94-XX, 94Axx, 62-XX, 62Mxx, 60-XX, 60Gxx 1. Introduction In spite of their practical importance in many fields of science and technology, apparently there is no universal and well-accepted definition of outlying data. This may have resulted from the wide range of fields (statistics, data mining, signal pr