augment: Augmentation of simple_brix model

In kbrix/brixtools: Some statistical methods for non-life insurance

Description

Augments simple_brix model with unobserved claims (losses smaller than the deductibles) for claims model distribution testing.

Usage

augment(object)

Arguments

object

object of class "brix_simple", see brix_simple.

Details

Using the Poisson distribution for augmentation of unobserved claims. (Might implement a Geometric method).

Value

Returns an object of class "augment". Using the print command displays test results from both the Shaprio-Wilks Normality test amd the Kolmogorov-Smirnov one-sample test for the standard normal transformed augmented data. See shapiro.test and ks.test. Using the plot command displays standard normal Q-Q plots for the standard normal transformed augmented data by default.

Examples

b <- brix_simple(optimizer = optim,
                 fn = normal_poisson_log_likelihood,
                 par = c(1, 1, 1),
                 control = list(fnscale = -1, maxit = 1e6),
                 hessian = TRUE,
                 X = normal_poisson_x, N = normal_poisson_n)

a <- augment(object = b)
print(a)
plot(a)
plot(a, augment = FALSE)

kbrix/brixtools documentation built on May 29, 2019, 2:25 p.m.