test.smoother: Test how well a smoother can filter noise from data

In maiers94/rpatrec: Recognising Visual Charting Patterns in Time Series Data

Description

Use this function to find out about the rate of successful recognition of a simple HS pattern for any smoothing function of your choice. Speciffy the noise and other testing conditions.

Usage

test.smoother(n = 1, m = 5, incr = 1, max = 20, smoother,
  pattern = TRUE, ntype = "white", ...)

Arguments

n	number of runs
m	number of runs per level of noise
incr	value by which the error is increased in each turn
max	max number of times the error is increased
smoother	Function with pre-defined inputs, so that only the parameter input is left to be defined
pattern	Check whether pattern was recognised. If FALSE only the correct position of extrema is checked.
ntype	Noise type. See noise function for details.
...	other parameters the smoother requires

Details

For an overview of the package capabilities, click here rpatrec. Note that this function may be extremely computationally demanding.

Value

Vector of recognition rates for specified levels of noise

Examples

## Not run: 
#Test the kernel regression smoother
a <- test.smoother(n=5,m=10,incr=0.5,max=50,smoother = kernel,bandwidth=1)
#Plot the result
plot(a,type='l')

## End(Not run)

maiers94/rpatrec documentation built on May 21, 2019, 11:06 a.m.