| ci {gmodels} | R Documentation |
Compute and display confidence intervals for model
estimates. Methods are provided for the mean of a numeric vector
ci.default, the probability of a binomial vector
ci.binom, and for lm, and lme
objects are
provided.
ci(x, confidence=0.95, alpha=1 - confidence, ...) ## S3 method for class 'numeric' ci(x, confidence=0.95, alpha=1-confidence, na.rm=FALSE, ...) ## S3 method for class 'binom' ci(x, confidence=0.95, alpha=1-confidence, ...) ## S3 method for class 'lm' ci(x, confidence=0.95, alpha=1-confidence, ...) ## S3 method for class 'lme' ci(x, confidence=0.95, alpha=1-confidence, ...) ## S3 method for class 'estimable' ci(x, confidence=0.95, alpha=1-confidence, ...)
x |
object from which to compute confidence intervals. |
confidence |
confidence level. Defaults to 0.95. |
alpha |
type one error rate. Defaults to 1.0- |
na.rm |
boolean indicating whether missing values should be
removed. Defaults to |
... |
Arguments for methods |
ci.binom computes binomial confidence intervals using the
Clopper-Pearson 'exact' method based on the binomial quantile
function. Due to the discrete nature of the binomial distribution,
this interval is conservative.
vector or matrix with one row per model parameter and elements/columns
Estimate, CI lower, CI upper, Std. Error,
DF (for lme objects only), and p-value.
Gregory R. Warnes greg@warnes.net
# mean and confidence interval ci( rnorm(10) ) # binomial proportion and exact confidence interval b <- rbinom( prob=0.75, size=1, n=20 ) ci.binom(b) # direct call class(b) <- 'binom' ci(b) # indirect call # confidence intervals for regression parameteres data(state) reg <- lm(Area ~ Population, data=as.data.frame(state.x77)) ci(reg) # mer example library(nlme) Orthodont$AgeGroup <- gtools::quantcut(Orthodont$age) fm2 <- lme(distance ~ Sex + AgeGroup, data = Orthodont,random=~1|Subject) ci(fm2)