Tail Extrapolation from Truncated Cumulative Event Curves

Fits parametric distributions to right-truncated event times using maximum likelihood estimation. Estimates the eventual total number of events, observed completeness, empirical likelihood-envelope bounds, and times to reach specified percentages of the eventual total.

tailcdf

tailcdf fits parametric distributions to right-truncated event times to estimate the unobserved tail of a cumulative event process.

It estimates:

• Eventual total number of events
• Percentage of events already observed
• Empirical likelihood-envelope bounds
• Times to reach 60%, 70%, 80%, and 90% completeness

The bounds are not conventional confidence intervals.

Installation

After CRAN release:

install.packages("tailcdf")

Example

library(tailcdf)

set.seed(123)
x <- rweibull(100, shape = 2, scale = 10)
x <- x[x <= 12]

fit <- tailfit(x, dists = "weibull")
print(fit)
plot(fit)

Input template

A customizable input template is included in the package. It provides distribution selection, bound-width adjustment, and completeness reporting options.