The method of least squares in a problem that does not have a solution

Бесплатный доступ

In this article, the author examines the potential of using the least squares method in order to predict the values of random variables that are in linear dependence. The author also reveals the inconsistency of the least absolute deviations method for solving problems of this kind and substantiates the idea that with a sufficient number of observations of two correlated variables, the law of causal relationship between them can be revealed as reliably as possible. For this purpose, in particular, an example of solving the problem of drawing a straight line through a cloud of points that do not lie on one straight line is given.

Еще

Forecast, causal law, regressor, resulting value, linear dependence, deviation

Короткий адрес: https://sciup.org/170194966

IDR: 170194966   |   DOI: 10.24412/2500-1000-2022-7-2-167-171

Статья научная