accidents: Accidents Dataset

In mixComp: Estimation of Order of Mixture Distributions

Description

Number of accidents incurred by 414 machinists over a period of three months from Karlis and Xekalaki (1999).

Usage

data(accidents)

Format

A data frame with 414 observations on 1 variable. Replicates are generated to reflect the number of accidents n incurred by michinists over a tree-month period (n = 0, 2, ..., 8). As there are 296 machinists that had no accidents, 0 appears 296 times in the data, as there are 74 machinists that had one accident, 1 appears 74 times in the data, etc.

Source

Karlis, D., Xekalaki, E. (1999) On Testing for the Number of Components in a Mixed Poisson Model. Annals of the Institute of Statistical Mathematics 51, 149-162.

Examples

data(accidents)

# convert the data to vector:
accidents.obs <- unlist(accidents)

# generate MLE function:
MLE.pois <- function(dat) mean(dat)

# generate function needed for estimating the j^th moment of the
# mixing distribution via Hankel.method "explicit"

explicit.pois <- function(dat, j){
  mat <- matrix(dat, nrow = length(dat), ncol = j) -
         matrix(0:(j-1), nrow = length(dat), ncol = j, byrow = TRUE)
  return(mean(apply(mat, 1, prod)))
}

# construct a 'datMix' object:
accidents.dM <- datMix(accidents.obs, dist = "pois", discrete = TRUE,
                      Hankel.method = "explicit",
                      Hankel.function = explicit.pois,
                      theta.bound.list = list(lambda = c(0, Inf)),
                      MLE.function = MLE.pois)

# define the penalty:
pen <- function(j, n) j * log(n)

## complexity estimation:
set.seed(0)
res <- paramHankel(accidents.dM, j.max = 5, B = 1000, ql = 0.025, qu = 0.975)
                               
# plot the results:
plot(res, breaks = 8, ylim = c(0, 0.8))

mixComp documentation built on Feb. 25, 2021, 5:07 p.m.