ftnonpar (version 0.1-88)

smdenreg: Piecewise monotone density estimation with smooth taut strings

Description

Applies the smooth taut string method to one-dimensional data.

Usage

smdenreg(x, verbose = FALSE, bandwidth=-1, maxkuipnr=19,asympbounds=FALSE, 
squeezing.factor=0.9, firstlambda=10,smqeps=1/length(x),fsign=double(0),
gensign=TRUE,...)

Arguments

x

observed values

verbose

logical, if T progress (for each iteration) is illustrated grahically

bandwidth

if set to a positive value the specified bandwidth is used instead of the automatic criterion based on generalized Kuiper metrics.

maxkuipnr

The order of the highest generalized Kuiper metric used for the automatic choice of the bandwidth

asympbounds

If set to T asymptotic bounds derived from a Brownian Bridge are used for the Kuiper criterion. Otherwise simulated bounds for various sample sizes are interpolated for the size of the data x

squeezing.factor

The amount of decrement applied to the bandwidthes

firstlambda

Initial value of lambda's for global squeezing.

smqeps

Distance between the (equally-spaced) time points.

fsign

Monotonicity constraints, vector of size n-1 of -1,0 and 1's. If fsign[i]==1, then fhat[i+1]>= fhat[i]. If fsign[i]==-1, then fhat[i+1]

gensign

If TRUE the taut string method is used to automatically produce suitable monotonicity constraints.

...

Passed to the plot command if verbose=T.

Value

x

The sorted data

y

values of the density approximation between the observations

nmax

Number of local extreme values

trans

taut string at the observations, should look like uniform noise

References

Kovac, A. (2006) Smooth functions and local extreme values. Technical Report

See Also

pmreg,l1pmreg,pmspec

Examples

y <- rclaw(500)
hist(y,col="lightgrey",40,freq=FALSE)
lines(smdenreg(y),col="red")