Abstract:Concerning bivariate least squares linear regression, the classical approach pursued for functional models in earlier attempts is reviewed using a new formalism in terms of deviation (matrix) traces. Within the framework of classical error models, the dependent variable relates to the independent variable according to the usual additive model. Linear models of regression lines are considered in the general case of correlated errors in X and in Y for heteroscedastic data. The special case of (C) generalized orthogonal regression is considered in detail together with well known subcases. In the limit of homoscedastic data, the results determined for functional models are compared with their counterparts related to extreme structural models. While regression line slope and intercept estimators for functional and structural models necessarily coincide, the contrary holds for related variance estimators even if the residuals obey a Gaussian distribution, with a single exception. An example of astronomical application is considered, concerning the [O/H]-[Fe/H] empirical relations deduced from five samples related to different stars and/or different methods of oxygen abundance determination. For selected samples and assigned methods, different regression models yield consistent results within the errors for both heteroscedastic and homoscedastic data. Conversely, samples related to different methods produce discrepant results, due to the presence of (still undetected) systematic errors, which implies no definitive statement can be made at present. A comparison is also made between different expressions of regression line slope and intercept variance estimators, where fractional discrepancies are found to be not exceeding a few percent, which grows up to about 20% in presence of large dispersion data.

Use and abuse of variance models in regression

Regressions with Berkson errors in covariates - A nonparametric approach

Estimation of the Variance Function in Heteroscedastic Linear Regression Models

Iterative weighted estimation based on variance modelling in linear regression models

Variance function estimation in high-dimensions

Bayesian variable selection in linear regression models with instrumental variables

Experimental Designs for Heteroskedastic Variance

Model Averaging Estimation for Nonparametric Varying-Coefficient Models with Multiplicative Heteroscedasticity

Mean and variance estimation in high-dimensional heteroscedastic models with non-convex penalties

Bivariate least squares linear regression: towards a unified analytic formalism

Nonlinear Regression Models With Single-Index Heteroscedasticity

Model averaging in a multiplicative heteroscedastic model

Models as Approximations I: Consequences Illustrated with Linear Regression

Variance Component Testing in Semiparametric Mixed Models

Separating biological variance from noise by applying EM algorithm to modified General Linear Model

Predicting uptake and elimination kinetics of chemicals in invertebrates: a technical note on residual variance modeling

A compromise criterion for weighted least squares estimates

Variance-based sensitivity analysis for weighting estimators result in more informative bounds

Error Variance Estimation in Partially Linear Varying Coefficient Models

Generalized varying coefficient partially linear measurement errors models

Heteroscedasticity Check in Nonlinear Semiparametric Models Based on Nonparametric Variance Function