Login for students

Login for employees

Publication detail

Problem of Regression Analysis and its Unconventional solution
Authors: Linda Bohdan | Kubanová Jana
Year: 2007
Type of publication: článek ve sborníku
Name of source: ISBIS 2007
Publisher name: University of Azores
Place: Azores, Portugal
Page from-to: 144
Titles:
Language Name Abstract Keywords
cze Problém regresní analýzy a jeho nekonvenční řešení Využití lineárního regresního modelu při statistických analýzách není zcela universální, musí být splněna řada předpokladů.Náhodné chyby se musí řídit normálním rozložení pravděpodobností a jejich rozptyly musí být konstantní.Pokud tyto předpoklady splněny nejsou, je obtížné odvodit rozdělení pravděpodobností odhadu regresních koeficientů. Metodou, která pomáhá řešit tuto situaci, je bootstrapový resampling.Aplikace této metody je popsána v článku. lineární regresní model, odhady parametrů, bootstrap, resampling
eng Problem of Regression Analysis and its Unconventional solution The linear regression model is one of the most important methods of statistical inference having wide practical use. This method is unfortunately used in the universal way without examining its assumptions. The essential assumption is that random errors have the normal distribution and their variances are constant values. The properties of the estimates obtained in this way are well known. In case when random errors are not normally distributed or their variances are not constant values (i.e., it changes depending on values of independent variable) it is seldom possible to derive the properties of distribution of regression coefficients estimates. The asymptotical properties of these estimates based on central limit theorem are usually assigned. They correspond with reality only in cases of large random samples. In reality we have small sample sizes so that the results obtained by these methods are not reliable enough. Other approximations can be used to determine the properties of estimates in some cases but these methods often require the complicated mathematical apparatus. One of methods that can help us to solve mentioned problem is the bootstrap resampling. It is advantageous to use these bootstrap methods in such cases. Linear regression model, parameters estimates, bootstrap, resampling

Map

en.mapy.cz

Services

Popular Links

© 2023 University of Pardubice, Studentská 95, 532 10 Pardubice 2
Phone: +420 466 036 111, 466 036 112, 466 036 113
Webmaster, RSS
Data box ID: f5vj9hu
Accessibility Statement
CC BY-NC-ND 4.0Creative CommonsAttributionNonCommercialNoDerivatives