Estimate linear combinations from a regression model

Usage

estimate(fit, lambda, alpha = 0.05, t.test = inherits(fit, "lm"), trans)

Arguments

fit: A model with coef() and vcov() methods
lambda: Linear combination weights (numeric vector with length equal to number of coefficients in fit)
alpha: Error rate for confidence interval (default: 0.05)
t.test: Use t distribution for inference? (default: TRUE only if model is linear)
trans: Apply a transformation function to the results? (function)

Examples

# Linear regression
fit <- lm(Ozone ~ Wind + Temp + Solar.R, airquality)
estimate(fit, c(0, 1, -1, 5))
#>      Estimate         Lower         Upper            SE             t 
#> -4.686581e+00 -5.871574e+00 -3.501589e+00  5.977614e-01 -7.840221e+00 
#>             p 
#>  3.561935e-12 

# Logistic regression
DF <- as.data.frame(Titanic)
DF <- DF[rep(1:nrow(DF), DF$Freq),]
fit <- glm(Survived ~ Class + Sex + Age, DF, family=binomial)
estimate(fit, c(0, 1, -1, 0, 0, 0), trans=exp)
#>     Estimate        Lower        Upper           SE            t            p 
#> 2.137565e+00 1.512603e+00 3.020742e+00           NA 4.307663e+00 1.722593e-05