[Rsiena-help] selecting certain actors

[Ruth M. Ripley]

Dear Maarten,

Sorry. No. Siena does not exclude 0's. But I should have just asked what
the scores are: a behavior variable or the strength of a relationship or a
covariate. I assumed the second, probably wrongly!  If a covariate or
behavior variable you just need to change the definition of use:

use <- score %in% c(1,2,3,4)

I suggested na.omit and NA's to select rows or columns from a network
matrix, but that may not be your problem!

Regards,

Ruth
--
Ruth M. Ripley,                         Email:ruth at stats.ox.ac.uk
Dept. of Statistics,                    http://www.stats.ox.ac.uk/~ruth/
University of Oxford,                   Tel:   01865 282851
1 South Parks Road, Oxford OX1 3TG, UK  Fax:   01865 272595


On Mon, 8 Mar 2010, Maarten Van Zalk wrote:

> Dear ruth,
>
> to clarify your response on changing to 0: do you mean that by making
> the other actor's scores zero, Siena will exclude them (I assume not)?
> what does 'net' refer to?
>
> On excluding other actors:
>
> I do not understand what the command "net <- na.omit(net)" does; does
> it excludes all 0's?? or all missings? Is it after the first command
> (net[!net %in% c(1,2,3,4)] <- 0) you use this in order to make all 0's
> missing?
>
> Kind regards,
>
> Maarten
>
>
>
>
>
> On Mar 8, 2010, at 16:01 , Ruth M. Ripley wrote:
>
> > Dear Maarten,
> >
> > Do you want to change the others to 0? Or exclude them?  To do the
> > former:
> >
> > net[!net %in% c(1,2,3,4)] <- 0
> >
> > To exclude them the easiest way is to make them missing and use
> > na.omit:
> >
> > net <- na.omit(net)
> >
> > You will need to alter the real missings to something else, though eg
> >
> > net[is.na(net)] <- 999
> >
> > then do the other two lines and then
> >
> > net[net==999] <- NA
> >
> > Regards,
> >
> > Ruth
> > --
> > Ruth M. Ripley,                         Email:ruth at stats.ox.ac.uk
> > Dept. of Statistics,                    http://www.stats.ox.ac.uk/~ruth/
> > University of Oxford,                   Tel:   01865 282851
> > 1 South Parks Road, Oxford OX1 3TG, UK  Fax:   01865 272595
> >
> >
> > On Mon, 8 Mar 2010, Maarten Van Zalk wrote:
> >
> >> Dear Ruth,
> >>
> >> a follow-up question on using specific actors: is it possible to
> >> select a
> >> range of scores? so for example the scores 1 until 4?
> >>
> >> Thanks for your help yet again!
> >>
> >> Maarten
> >>
> >> On Mon, Feb 15, 2010 at 12:59 PM, Ruth M. Ripley
> >> <ruth at stats.ox.ac.uk>wrote:
> >>
> >>> Dear Maarten,
> >>>
> >>> There are no special facilities: you would need to extract the
> >>> subsets of
> >>> the matrices yourself before creating the Siena network object.
> >>> Using the
> >>> command line, after reading in some data:
> >>>
> >>> friend.data.w1 <- as.matrix(read.table("s50-network1.dat"))
> >>> friend.data.w2 <- as.matrix(read.table("s50-network2.dat"))
> >>> drink <- as.matrix(read.table("s50-alcohol.dat"))
> >>> smoke <- as.matrix(read.table("s50-smoke.dat"))
> >>>
> >>> To select the actors with the first column of smoke equal to 1, say,
> >>>
> >>> use <- smoke[, 1] == 1
> >>>
> >>> friendship <- sienaNet(array(c(friend.data.w1[use, use],
> >>>      friend.data.w2[use, use]), dim=c(38, 38, 2)))
> >>>
> >>> alcohol <- sienaNet(drink[use, ], type="behavior")
> >>>
> >>> To do this via the gui, you would need to do the subsetting
> >>> outside R.
> >>>
> >>> Regards,
> >>>
> >>> Ruth
> >>>
> >>> PS I have copied this message to the rsiena-help mailing list. I
> >>> see two
> >>> benefits to using the lists: others might find the information
> >>> useful too,
> >>> and if I am busy someone else might be able to help!
> >>> --
> >>> Ruth M. Ripley,                         Email:ruth at stats.ox.ac.uk<Email%3Aruth at stats.ox.ac.uk
> >>> >
> >>> Dept. of Statistics,                    http://www.stats.ox.ac.uk/~ruth/
> >>> University of Oxford,                   Tel:   01865 282851
> >>> 1 South Parks Road, Oxford OX1 3TG, UK  Fax:   01865 272595
> >>>
> >>>
> >>> On Mon, 15 Feb 2010, Maarten Van Zalk wrote:
> >>>
> >>>> Dear Ruth,
> >>>>
> >>>> Is it possible in RSiena to select certain actors for the
> >>>> analysis? So
> >>>> if you have a variable X you want to use to select individuals (for
> >>>> example, all individuals with a score of 1), Can you select only
> >>>> these
> >>>> individuals?
> >>>>
> >>>> Thanks for your help!
> >>>>
> >>>> Kind regards,
> >>>>
> >>>> Maarten Van Zalk
> >>>>
> >>>
> >>>
> >>
> >
>
>