For each cluster, this function determines a ring of equal number of observations and/or equal radius and calculates several indicators from observations located inside that ring.
# S4 method for prevR
rings(object, N = seq(100, 500, 50), R = Inf,
progression = TRUE)
object of class prevR
.
minimum number of observations.
maximum rings radius (in kilometers if coordinates in decimal degrees, in the unit of the projection otherwise).
show a progress bar?
Return object
with the slot rings
completed for each couple (N,R).
Each entry is composed of 3 elements: N
, minimum number of observations per ring; R
,
maximum radius of rings and estimates
, a data frame with the following variables:
"id" cluster ID.
"r.pos" number of positive cases inside the ring.
"r.n" number of valid observations inside the ring.
"r.prev" observed prevalence (in %) inside the ring (r.pos/r.n).
"r.radius" ring radius (in kilometers if coordinates in decimal degrees, in the unit of the projection otherwise).
"r.clusters" number of clusters located inside the ring.
"r.wpos" (optional) sum of weights of positive cases inside the ring.
"r.wn" (optional) sum of weights of valid observations inside the ring.
"r.wprev" (optional) weighted observed prevalence (in %) inside the ring (r.wpos/r.wn).
Note: the list rings
is named, the name of each element is NN_value.RR_value,
for example N300.RInf.
Note 2: r.wpos, r.wn and r.wprev are calculated only if the slot clusters
of object
contains weighted data.
For each ligne of the data frame clusters
of object
, rings
determines
a ring, centred on the cluster. It could be:
rings of eaqul number of observations if N
is finite and R=Inf
;
rings of equal radius if N=Inf
and R
is finite;
a combination of both (see below) if N
and R
are finite.
For rings of equal number of observations, rings
selects the smallest
ring containing at least N
valid observations.
For rings of equal radius, rings
selects all clusters located at a lower
distance than R
from the central cluster. For combination of both, rings
calculates firts the ring with the minimum number of observations and test if its radius is lower
than R
or not. If so, the ring is kept, otherwise the ring of maximum radius is calculated.
Different series of rings could be simultaneoulsy calculated by providing different values for N
and R
. rings
will calculate rings corresponding to each couple (N,R).
Larmarange Joseph, Vallo Roselyne, Yaro Seydou, Msellati Philippe and Meda Nicolas (2011) "Methods for mapping regional trends of HIV prevalence from Demographic and Health Surveys (DHS)", Cybergeo : European Journal of Geography, no 558, http://cybergeo.revues.org/24606, DOI: 10.4000/cybergeo.24606.
Larmarange Joseph (2007) Pr<U+00E9>valences du VIH en Afrique : validit<U+00E9> d'une mesure, PhD thesis in demography, directed by Beno<U+00EE>t Ferry, universit<U+00E9> Paris Descartes, http://tel.archives-ouvertes.fr/tel-00320283.
# NOT RUN {
print(fdhs)
dhs <- rings(fdhs,N=c(100,200,300,400,500))
print(dhs)
# }
Run the code above in your browser using DataLab