Linganisha mbinu

	Weighted Least Squares (WLS)	Ordinary Least Squares (OLS)
Nyanja	Takwimu	Takwimu
Familia	Regression model	Regression model
Mwaka wa asili	1935	1805
Mwanzilishi	Alexander Craig Aitken	Adrien-Marie Legendre (1805); Carl Friedrich Gauss (1809)
Aina	Weighted linear estimator	Linear parameter estimation
Chanzo asilia	Aitken, A. C. (1935). IV.—On least squares and linear combination of observations. Proceedings of the Royal Society of Edinburgh, 55, 42–48.	Legendre, A.-M. (1805). Nouvelles méthodes pour la détermination des orbites des comètes. Firmin Didot, Paris. [Appendix: Sur la Méthode des moindres quarrés, pp. 72–80.]
Majina mbadala	WLS, weighted regression, heteroscedasticity-corrected OLS, variance-weighted least squares	OLS, OLS regression, linear least squares, classical linear regression
Zinazohusiana	3	8
Muhtasari	Weighted Least Squares is a generalization of Ordinary Least Squares (OLS) regression that assigns each observation a weight inversely proportional to its error variance, thereby down-weighting high-variance data points and up-weighting precise ones. Introduced in its general matrix form by Alexander Craig Aitken in 1935, WLS is the canonical remedy when heteroscedasticity is present and the error variance structure is known or can be reliably estimated.	Ordinary Least Squares (OLS) is the canonical method for estimating the parameters of a linear regression model by minimizing the sum of squared differences between observed and predicted values. First published by Adrien-Marie Legendre in 1805 and independently developed by Carl Friedrich Gauss (who claimed priority from 1795), OLS is provably optimal under the Gauss-Markov theorem: given its assumptions, it yields the Best Linear Unbiased Estimator (BLUE) of the regression coefficients.