stat.poisson.adj: Compute Poisson test statistic

In jfrench/smerc: Statistical Methods for Regional Counts

Compute Poisson test statistic

Description

Efficiently compute a vector of Poisson test statistics. This function does no sanity checking. We encourage you to compare the results against scan.stat for correctness. stat_poisson_adj is a C++ version implementation of the code and should be faster. stat_binomial_adj is a C++ version implementation of stat.binom.

Usage

stat.poisson.adj(
  yin,
  ty,
  logein,
  logeout,
  a = 0,
  pen = 1,
  min.cases = 2,
  return.max = FALSE
)

stat_poisson_adj(
  yin,
  ty,
  logein,
  logeout,
  a = 0,
  pen = 1,
  min.cases = 2,
  return.max = FALSE
)

stat_binom_adj(
  yin,
  ty,
  popin,
  popout,
  logpopin,
  logpopout,
  tpop,
  min.cases = 2,
  return.max = FALSE
)

Arguments

yin	The total number of cases in the zone.
ty	The total number of cases in the study area.
logein	The log of the expected number of cases in each candidate zone.
logeout	The log of the expected number of cases outside of each candidate zone.
a	A tuning parameter for the adjusted log-likelihood ratio. See details.
pen	The eccentricity penalty for each candidate zone.
min.cases	The minimum number of cases required for a cluster. The default is 2.
return.max	A logical value. Default is FALSE. If TRUE, then only the maximum statistic is returned.
popin	The total population in the zone.
popout	The population outside the zone. This should be tpop - popin and is computed automatically if not provided.
logpopin	The log of the population in each candidate zone.
logpopout	The log of the population outside of each candidate zone.
tpop	The total population in the study area.

Value

A numeric vector.

Examples

data(nydf)
coords <- with(nydf, cbind(longitude, latitude))
enn <- elliptic.nn(coords, nydf$pop, ubpop = 0.015)
cases <- floor(nydf$cases)
ty <- sum(cases)
ex <- ty / sum(nydf$pop) * nydf$pop
yin <- nn.cumsum(enn$nn, cases)
ein <- nn.cumsum(enn$nn, ex)
logein <- log(ein)
logeout <- log(ty - ein)
pen <- elliptic.penalty(a = 0.5, enn$shape_all)
stat.poisson.adj(yin, ty, logein, logeout,
  a = 0.5, pen = pen, return.max = TRUE
)

jfrench/smerc documentation built on Oct. 27, 2024, 5:13 p.m.