Skip to main content
Log in

Calculo practico de la regresion en mediana

  • Published:
Trabajos de Estadistica

Summary

This article explains a metrod to find out the regression line or plane by median similar to linear programing method

The regression line is y=a+bx if\(z = \sum {\left| {y_i - a - bx_i } \right|} \) is minimum.

We calculate the intersection of two straight linesi·e:

$$y_i = a + bx_i i = 1 \cdot 2$$

and we shall verify if the point found out produces a minimum inz. If that did not happen, changing sistematicaly the values ora andb we should have the solution.

Also, the regression plane is

$$z = a + bx + cy if Z = \sum {\left| {z_i - a - bx_i - cy_i } \right|} $$

is minimum.

We give, too, the disposition of data in order to make the shortest calculations, as you can see by the two examples.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Similar content being viewed by others

Authors

Rights and permissions

Reprints and permissions

About this article

Cite this article

Béjar, J. Calculo practico de la regresion en mediana. Trabajos de Estadistica 8, 157–173 (1957). https://doi.org/10.1007/BF03006468

Download citation

  • Issue Date:

  • DOI: https://doi.org/10.1007/BF03006468

Navigation