Year: 1994
Type: erratum
Abstract: Measures of variability of the least-squares estimator 9 are essential to assess the qual ity of the estimation. In nonlinear regression, an accurate of the covariance matrix of 0 is difficult to obtain [4]. In this paper, a second-order of the entropy of the distribution of 8 is proposed, which is only slightly more complicated than the widely used bias of Box [3]. It is based on the flat or saddle-point approximation of the density of 6. The neglected terms are of order 0(tr*), while the clas sical first order neglects terms of order 0(ff2). Various illustrative examples are presented, including the use of the approximate entropy as a criterion for experimental design.
Authors: Luc Pronzato, Andrej Pázman
Language: English
Subfield: Applied Mathematics
Field: Mathematics
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W2396010748