mvi {gamlss.data}

The motor vehicle insurance data

Description

The motor vehicle insurance data are motor vehicle insurance policies. mvi is a sample of 2000 observations from mviBig which has 67143 observartions

Usage

data(mvi)
data(mviBig)

Format

Two data frames with 2000 or 67143 observations on the following 14 variables.

retval

a numeric vector showing the value of the vehicle

whetherclm

a numeric vector showing whether a claim is made, 0 no claim, 1 at least one claim

numclaims

a nuneric vactor showing the number of claims

claimcst0

a numeric vector showing the total amount of claim, i.e. for numclaims=0 is zero.

vehmake

a factor showing the make of the car with levels BMW DAEWOO FORD MITSUBISHI

vehbody

a factor showing the type of the cat, with levels BUS CONT COUPE HACK HDTOP HRSE MCARA MIBUS PANVN RDSTR SEDAN STNWG TRUCK UTE

vehage

a numeric vector showing the age of the car

gender

a factor showing the gender of the policy holder with levels F M

area

a factor showing the Area of residence of the policy holder with levels A B C D E F

agecat

a factor showing the age band of the policy holder with levels 1 2 3 4 5 6 one is youngest

exposure

a numeric vector showing the time of exposure with values from zero to one

Details

The motor vehicle insurance data are motor vehicle insurance policies from an insurance company over a twelve-month period in 2004-05. The original data are 67143 observation but here we also include a random sample of 2000.

References

Heller, G. Stasinopoulos M and Rigby R.A. (2006) The zero-adjusted Inverse Gaussian distribution as a model for insurance claims. in Proceedings of the 21th International Workshop on Statistial Modelling, eds J. Hinde, J. Einbeck and J. Newell, pp 226-233, Galway, Ireland.

Heller G. Z., Stasinopoulos M.D., Rigby R. A. and de Jong P. (2007) Mean and dispersion modeling for policy claims costs. To be published in the Scandinavian Actuarial Journal.

Examples

data(mvi)
## a histogram of claims with fitted gamma disteibution
## library(gamlss)
## with(mvi, histDist(claimcst0[whetherclm==1&claimcst0<15000], family=GA, main="Claims"))