aboutsummaryrefslogtreecommitdiff
path: root/R
diff options
context:
space:
mode:
authorJohannes Ranke <jranke@uni-bremen.de>2019-10-28 13:23:40 +0100
committerJohannes Ranke <jranke@uni-bremen.de>2019-10-28 13:23:40 +0100
commitbd761d879a95872f82e6d8f893634a61e122a938 (patch)
tree0b66fa5064eb5bd9e0014de18f433c2692fcbc00 /R
parent6e5630a0df7e857697ff2ce4730a5f7f45b67377 (diff)
Fix the cutoff for likelihood based intervals
The cutoff now matches what is given by Venzon and Moolgavkar (1988). Also, confidence intervals closely match intervals obtained with stats4::confint in the test case where an stats4::mle object is created from the likelihood function in one test case. Static documentation rebuilt by pkgdown
Diffstat (limited to 'R')
-rw-r--r--R/confint.mkinfit.R20
1 files changed, 16 insertions, 4 deletions
diff --git a/R/confint.mkinfit.R b/R/confint.mkinfit.R
index 58a3c8db..1ed689a9 100644
--- a/R/confint.mkinfit.R
+++ b/R/confint.mkinfit.R
@@ -1,10 +1,17 @@
#' Confidence intervals for parameters of mkinfit objects
#'
+#' The default method 'profile' is based on the profile likelihood for each
+#' parameter. The method uses two nested optimisations. The speed of the method
+#' could likely be improved by using the method of Venzon and Moolgavkar (1988).
+#'
#' @param object An \code{\link{mkinfit}} object
#' @param parm A vector of names of the parameters which are to be given
#' confidence intervals. If missing, all parameters are considered.
#' @param level The confidence level required
#' @param alpha The allowed error probability, overrides 'level' if specified.
+#' @param cutoff Possibility to specify an alternative cutoff for the difference
+#' in the log-likelihoods at the confidence boundary. Specifying an explicit
+#' cutoff value overrides arguments 'level' and 'alpha'
#' @param method The 'profile' method searches the parameter space for the
#' cutoff of the confidence intervals by means of a likelihood ratio test.
#' The 'quadratic' method approximates the likelihood function at the
@@ -25,6 +32,9 @@
#' @importFrom stats qnorm
#' @references Pawitan Y (2013) In all likelihood - Statistical modelling and
#' inference using likelihood. Clarendon Press, Oxford.
+#' Venzon DJ and Moolgavkar SH (1988) A Method for Computing
+#' Profile-Likelihood Based Confidence Intervals, Applied Statistics, 37,
+#' 87–94.
#' @examples
#' f <- mkinfit("SFO", FOCUS_2006_C, quiet = TRUE)
#' confint(f, method = "quadratic")
@@ -33,7 +43,7 @@
#' }
#' @export
confint.mkinfit <- function(object, parm,
- level = 0.95, alpha = 1 - level,
+ level = 0.95, alpha = 1 - level, cutoff,
method = c("profile", "quadratic"),
transformed = TRUE, backtransform = TRUE,
distribution = c("student_t", "normal"), quiet = FALSE, ...)
@@ -100,15 +110,17 @@ confint.mkinfit <- function(object, parm,
}
if (method == "profile") {
- message("Profiling the likelihood")
+ if (!quiet) message("Profiling the likelihood")
+
lci <- uci <- rep(NA, p)
names(lci) <- names(uci) <- return_pnames
profile_pnames <- if(missing(parm)) names(parms(object))
else parm
- # We do two-sided intervals based on the likelihood ratio
- cutoff <- 0.5 * qchisq(1 - (alpha / 2), 1)
+ if (missing(cutoff)) {
+ cutoff <- 0.5 * qchisq(1 - alpha, 1)
+ }
all_parms <- parms(object)

Contact - Imprint