Beyond Maximum Likelihood and Least Squares; Robust Methods

Autor:	Byron P. Roe
Rok vydání:	2020
Předmět:	Goodness of fit Statistics Deviance (statistics) Least trimmed squares Lack-of-fit sum of squares Estimating equations Generalized least squares M-estimator Mathematics Robust regression
Zdroj:	Probability and Statistics in the Physical Sciences ISBN: 9783030536930 Probability and Statistics in Experimental Physics ISBN: 9781441928955 Probability and Statistics in Experimental Physics ISBN: 9781475721881
Popis:	The methods discussed in Chapters 12, 13, and 16 are very powerful if the underlying distributions are close to normal. If the distributions of errors of the points on a curve, for example, are not normal and have long tails, then, as we noted in Chapter 13, estimates of goodness of fit may be seriously biased. The tests discussed in this chapter tend to be robust, with results which are independent of the particular distribution being tested.
Databáze:	OpenAIRE
Externí odkaz:	https://explore.openaire.eu/search/publication?articleId=doi_dedup___::f64a8c55bcc64da03d9bb556e61167ac https://doi.org/10.1007/978-3-030-53694-7_19 Zobrazit plný text záznamu