1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
|
1) we have had some problems, but they are solved now
---> tests/C-BUG-whimed.R
2) I think its wrong in principle to only work with *integer* weights
for 'whimed'.
Further note, that *because* of the integer weights, we need type
'long long' {because otherwise integer overflow kills everything}.
--> Change this `everywhere' and make wgt.himedian() an R function
NOTE: There is 'weighted.median()' in limma
and I have generalized it ---> inst/ex-funs.R
But see more general code (different 'tie' strategies;
weighted *quantile*s)
in /u/maechler/R/MM/STATISTICS/robust/weighted-median.R
3) Data Sets --- Valentin Todorov has several of Rousseeuw's
in his "rrov" package
(and promised me "the rest" when needed)
I don't like the *.x, *.y sub datasets:
They shouldn't be needed when use a *formula*
In his lts tests, he uses these "data sets from the literature":
(Note that 'stackloss' is already in "datasets") :
heart.x,heart.y, data(heart)
stars.x,stars.y, data(stars)
phosphor.x,phosphor.y, data(phosphor)
stack.x,stack.loss, data(stackloss)
coleman.x,coleman.y, data(coleman)
salinity.x,salinity.y, data(salinity)
aircraft.x,aircraft.y, data(aircraft)
delivery.x,delivery.y, data(delivery)
wood.x,wood.y, data(wood)
hbk.x,hbk.y, data(hbk)
|
