Tehničko veleučilište u Zagrebu · Zagreb

On estimating distances by means of L_p-norms

izvorni znanstveni rad

izvorni znanstveni rad

On estimating distances by means of L_p-norms

Vrsta prilog sa skupa (u zborniku)
Tip izvorni znanstveni rad
Godina 2003
Nadređena publikacija Proceedings of the 9th International Conference on Operational Research - KOI 2002
Status objavljeno

Sažetak

In problems of estimating the distance between objects in space (e.g. different cities or destinations in a geographic region), one can use a distance predicting function. Distance predicting functions transform the point coordinates of two objects into an estimate of the distance between them. We look at the weighted L_2-norm and generally the weighted L_p-norm (1<= p<=infinity) as the distance predicting functions. In the case of the weighted L_p-norm model (i.e. tau L_p models), it is necessary to estimate parameters tau and p. Then one copes with the problem of goodness-of-fit criteria for fitting the unknown parameters of a distance function to the data set. We consider this problem by an example of the set of data on road travel distances in a geographical area of the north-eastern part of Croatia.

Ključne riječi

estimating distances; L_p-norm model