insilico.fit {InSilicoVA} | R Documentation |
Implement InSilicoVA methods with more flexible customization
Description
This function implements InSilicoVA model. This is the lower level core function of InSilicoVA with more flexibility in customized input. For more detail of model specification, see the paper on https://arxiv.org/abs/1411.3042 and the default function insilico
.
Usage
insilico.fit(
data,
data.type = c("WHO2012", "WHO2016")[1],
sci = NULL,
isNumeric = FALSE,
updateCondProb = TRUE,
keepProbbase.level = TRUE,
CondProb = NULL,
CondProbNum = NULL,
datacheck = TRUE,
datacheck.missing = TRUE,
warning.write = FALSE,
directory = NULL,
external.sep = TRUE,
Nsim = 4000,
thin = 10,
burnin = 2000,
auto.length = TRUE,
conv.csmf = 0.02,
jump.scale = 0.1,
levels.prior = NULL,
levels.strength = 1,
trunc.min = 1e-04,
trunc.max = 0.9999,
subpop = NULL,
java_option = "-Xmx1g",
seed = 1,
phy.code = NULL,
phy.cat = NULL,
phy.unknown = NULL,
phy.external = NULL,
phy.debias = NULL,
exclude.impossible.cause = c("subset2", "subset", "all", "InterVA", "none")[1],
impossible.combination = NULL,
no.is.missing = FALSE,
customization.dev = FALSE,
Probbase_by_symp.dev = FALSE,
probbase.dev = NULL,
table.dev = NULL,
table.num.dev = NULL,
gstable.dev = NULL,
nlevel.dev = NULL,
indiv.CI = NULL,
groupcode = FALSE,
known_labels = NULL,
...
)
Arguments
data |
see |
data.type |
see |
sci |
see |
isNumeric |
see |
updateCondProb |
see |
keepProbbase.level |
see |
CondProb |
see |
CondProbNum |
see |
datacheck |
see |
datacheck.missing |
see |
warning.write |
see |
directory |
see |
external.sep |
see |
Nsim |
see |
thin |
see |
burnin |
see |
auto.length |
see |
conv.csmf |
see |
jump.scale |
see |
levels.prior |
see |
levels.strength |
see |
trunc.min |
see |
trunc.max |
see |
subpop |
see |
java_option |
see |
seed |
see |
phy.code |
see |
phy.cat |
see |
phy.unknown |
see |
phy.external |
see |
phy.debias |
see |
exclude.impossible.cause |
see |
impossible.combination |
see |
no.is.missing |
see |
customization.dev |
Logical indicator for customized variables |
Probbase_by_symp.dev |
Not tested yet. |
probbase.dev |
The customized conditional probabilities of symptoms given causes, which could be in a different format than InterVA default, but it should consist of |
table.dev |
The table of level names in |
table.num.dev |
The corresponding prior numerical values for each level in |
gstable.dev |
Table of gold standard causes for each death. Default to be NULL |
nlevel.dev |
number of levels in |
indiv.CI |
credible interval for individual probabilities |
groupcode |
logical indicator of including the group code in the output causes |
known_labels |
a data frame with two columns: the first column is the death ID and the second column is the known cause of death (need to match the cause list for the given data format). When it is provided for some causes, they will be used as partial labels in the input data. Any unmatched observations (unmatched by either ID or cause) will not contribute to partial labels. Default to be NULL |
... |
unused arguments |
Value
a insilico fit object, see see insilico
for more detail.
Author(s)
Zehang Li, Tyler McCormick, Sam Clark
Maintainer: Zehang Li <lizehang@uw.edu>
References
Tyler H. McCormick, Zehang R. Li, Clara Calvert, Amelia C.
Crampin, Kathleen Kahn and Samuel J. Clark(2014) Probabilistic
cause-of-death assignment using verbal autopsies,
https://arxiv.org/abs/1411.3042
Working paper no. 147, Center
for Statistics and the Social Sciences, University of Washington
See Also
plot.insilico
, summary.insilico