creatematrix: create a dominance matrix

In gobbios/EloRating: Animal Dominance Hierarchies by Elo Rating

create a dominance matrix

Description

create a dominance matrix from the underlying observed sequence

Usage

creatematrix(
  eloobject,
  daterange = NULL,
  drawmethod = "omit",
  onlyinteracting = FALSE,
  winners,
  losers,
  draw = NULL
)

Arguments

eloobject	output from elo.seq
daterange	character of length 2, date range to which the matrix should correspond (default from beginning to end of sequence)
drawmethod	character with the following options: "omit" = undecided interactions (draws/ties) are ignored (default) "0.5" = each undecided is counted half a win for each dyad member "1" = each undecided interaction is counted twice, i.e. as win for both individuals
onlyinteracting	logical, indicating whether all individuals that were present (default, TRUE) are shown in the matrix, or only those that were involved in an interaction in the specified date period. If no presence data was supplied to elo.seq, it is assumed that all individuals were present at all times
winners	vector of winners (see details)
losers	vector of losers (see details)
draw	logical vector (currently not doing anything)

Details

The function works with either the output of elo.seq, or with two vectors of winners and losers. If you use winner and loser vectors, their arguments need to be named, and also the remaining arguments (daterange= and onlyinteracting=) are ignored. The function does not yet allow to include draws if you supply winner/loser vectors. If you go via the elo.seq-route, the function can handle draws (via the drawmethod= argument).

Value

square matrix with dominance interactions (winner in rows, loser in columns)

Author(s)

Christof Neumann

Examples

data(adv)
# from winner/loser sequence directly
creatematrix(winners=adv$winner, losers=adv$loser)
# via an eloobject
SEQ <- elo.seq(winner=adv$winner, loser=adv$loser, Date=adv$Date)
# create dyadic matrix over the entire period of data collection
creatematrix(SEQ)
# limit to a subset of interactions
creatematrix(SEQ, daterange=c("2010-01-25", "2010-02-01"))
# limit to a subset of interactions and show only those IDs that were
# involved in at least one interaction
creatematrix(SEQ, daterange=c("2010-01-25", "2010-02-01"),
             onlyinteracting=TRUE)
             
# interactions restricted to single date
creatematrix(SEQ, daterange = c("2010-01-25", "2010-01-25"))

## dealing with undecided interactions
data(adv2)
SEQ <- elo.seq(winner=adv2$winner, loser=adv2$loser, Date=adv2$Date,
               draw=adv2$tie)
# omit ties/draws
creatematrix(SEQ)
# omit ties/draws
creatematrix(SEQ, drawmethod="0.5")
# omit ties/draws
creatematrix(SEQ, drawmethod="1")

gobbios/EloRating documentation built on July 19, 2024, 4:05 a.m.