rvmethod (version 0.1.2)

estimate_template: Estimate the Template Spectrum

Description

This function uses local quadratic regression to estimate the template spectrum from a collection of observed spectra from a star as described in Holzer et al. (2020). All observed spectra are assumed to be normalized. The bandwidth is chosen locally through generalized cross-validation. We strongly recommend using parallel computing for this function. Therefore, the cores argument has the default value of 19.

Usage

estimate_template(
  SPECTRA,
  min_wvl = NULL,
  max_wvl = NULL,
  bandwidth_bnds = c(0.017, 0.05),
  min_count = 100,
  cores = 19
)

Arguments

SPECTRA

a list of all observed spectra to use in estimating the template. Each observed spectrum should have the format of being a list with the following names (or a dataframe with the following columns): ``Wavelength" and ``Flux".

min_wvl

a number that indicates the minimum wavelength for the estimated template

max_wvl

a number that indicates the maximum wavelength for the estimated template

bandwidth_bnds

a vector of length 2 that gives the interval of bandwidth values (in the same units as the wavelength of the spectra) to be considered in the generalized cross-validation

min_count

the minimum number of data points required for local regression to be done on a given wavelength chunk

cores

the number of cores to parallelize over (if set to 1, no parallelizing is done)

Value

a list with the following elements:

Wavelength

the wavelength axis of the estimated template

Flux

the normalized flux of the estimated template

Chunk_bounds

a list of length 2 vectors that give the wavelength bounds for each chunk for which the smoothing was done on

Bandwidths

the bandwidths chosen for each of the chunks

Std_err

the standard errors of the estimated normalized flux that can be used for prediction confidence intervals

Examples

# NOT RUN {
data(spectra)
plot(spectra[[1]]$Wavelength, spectra[[1]]$Flux, col='gray', type='l')
for(spec in spectra){
 lines(spec$Wavelength, spec$Flux, col='gray')
}
tempest = estimate_template(spectra, cores = 1)
lines(tempest$Wavelength, tempest$Flux, col='red')
# }