bda (version 1.1.1-7)

histo: Draw histogram based on data with rounding errors

Description

Draw histogram based on data with rounding errors.

Usage

histo(x, m=NULL, alpha=0.05, binned=FALSE,just="center",scale=1.0)

Arguments

x

a numeric value or a vector.

m

number of bins.

alpha

Significance level. Default: 0.05.

binned

Have the data been binned. Default: FALSE.

just

How the data were binned: 'center' (default), 'left' or 'right'.

scale

Data can be rescaled. Default: 1 (no rescaling).

Value

• If 'scale'=100, the data will be rescaled by dividing all values by 100. For exam the birth weight are usually rounded to 100mg.

Details

The cross validation score will be return as 'Jh'.

See Also

edf, pcb.

Examples

x = rnorm(10000)
m=2:100
Jh = NULL
for(i in m){
  out=histo(x,i)
  Jh=c(Jh,out$Jh)
}
plot(Jh~m,type='l')

j = m[which(Jh==min(Jh))]
(out1=histo(x,j))
plot(out1)
lines(dnorm(sort(x))~sort(x))