I know you can compare Weibull distributions using the likelihood test but I was asking if there is a way to compare individual parameters like (shape and scale) between different distributions. A similar process is provided by Minitab I'd like to know if such approach is possible using R. A similar question was asked here but again the answer was talking about comparing the distribution as whole. I use the following code for comparing distributions:

compare_weibull <- function(x, y) {
    pooled <- fitdistr(x = c(x, y), densfun = "weibull")
    pooled <- pooled$log
    A <- fitdistr(x = x, densfun = "weibull")
    A <- A$log
    B <- MASS::fitdistr(x = y, densfun = "weibull")
    B <- B$log
    p <- pchisq(2 * ((A + B) - pooled), df = 2, lower.tail = FALSE)
    data.frame(stat = 2 * ((A + B) - pooled), p = p)

Thanks in advance.



Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.