Lad Regression

What is Lad Regression?
Secondly, how use it?

In this section we mention some tools to manage lad regression by Splus. We refer to the paper "LAD regression and non-parametric methods for detecting outliers and leverage points" by Susana Ferreira de Sá Faria and Giuseppe Melfi. In this paper we proposed a nonparametric method for detecting influencial observations. We provide here the Splus script (pointtotal8.s), and the data used in the paper.
pointtotal8.s
Hawk75by4.dat
telephone.dat
threevariables.dat
twovariables.dat
scottish.dat

In this section we refer to program Adaptive, available in the Dodge's homepage.

The program is written by Jan Picek, from the University of Liberec.

The three necessary files are

rq.s
adaptive.s
rq.obj

Files are also available on Jan Picek's web site.

The program requires at least version 4.5 of S-Plus.

After starting S-Plus, type

source("path\\rq.s")
source("path\\adaptive.s")

where path is the full name of the directory (c:\\dirname\\subdirname ...) where the files are located. These two files must be installed only once, but at the beginning of each session, you must type

dyn.load("path\\rq.obj")

Then the program is called in typing the followin command:

adaptive ("path\\filename", alpha=0.05, const=1.2, betatest=0, snkern=F, detail=F, diagn=F, tets=F, out="S", iter=150, toler=.Machine$double.eps^(1/2)))

It is necessary to give at least the first option.