CI_English: Confidence Intervals: Univariate and Multivariate.

In MVar: Multivariate Analysis

Confidence Intervals: Univariate and Multivariate.

Description

Computes univariate and multivariate confidence intervals for variables in a dataset.

Usage

CI(data, sigma = NA, conf = 0.95, type = "T")

Arguments

data	Data used to compute the confidence intervals.
sigma	Variance–covariance matrix; otherwise the CI will be based on the sample variance (default sigma = NA).
conf	Confidence level of the CI (default conf = 95%).
type	Type of interval: 'T' for Hotelling’s \(T^2\) or 'B' for Bonferroni.

Value

CIu	Univariate confidence interval with confidence level 'conf'.
CIm	Multivariate confidence interval with confidence level 'conf'.

Author(s)

Paulo Cesar Ossani

References

Ferreira, D. F. EstatistCIa Multivariada. 2a ed. revisada e ampliada. Lavras: Editora UFLA, 2011. 676 p.

Rencher, A. C. Methods of multivariate analysis. 2th. ed. New York: J.Wiley, 2002. 708 p.. 708 p.

See Also

Plot.CI

Examples

data(iris) # data set

# Interval with unknown population variance
res <- CI(data = iris[,1:4], sigma = NA, conf = 0.95, type = 'T') 
# res <- CI(data = iris[,1:4], sigma = NA, conf = 0.95, type = 'B') 

res$ciu # Univariate interval
res$cim # Multivariate interval


# Interval with known population variance
sig <- matrix(c(5.286, 0.942, 1.274, 4.516, # Example of known variance matrix
                0.942, 3.690, 0.730, 1.122,
                1.274, 0.730, 7.116, 1.296,
                4.516, 1.122, 1.296, 4.581), 
                nrow = 4, byrow = TRUE)
           
# res <- CI(data = iris[,1:4], sigma = sig, conf = 0.95, type = 'T')
res <- CI(data = iris[,1:4], sigma = sig, conf = 0.95, type = 'B')

res$ciu # Univariate interval
res$cim # Multivariate interval

MVar documentation built on June 22, 2026, 1:06 a.m.