Predicts sample properties from attenuated total reflection mid-infrared spectra

Functions to predict peat properties from attenuated total reflection (ATR) mid-infrared spectra. All functions below have been computed using various organic matter samples. For detailed information on the underlying prediction models, see the details section.

Usage

irp_holocellulose_content_2(
  x,
  ...,
  do_summary = FALSE,
  summary_function_mean = mean,
  summary_function_sd = stats::sd,
  return_as_list = FALSE,
  check_prediction_domain = "train"
)

irp_klason_lignin_content_2(
  x,
  ...,
  do_summary = FALSE,
  summary_function_mean = mean,
  summary_function_sd = stats::sd,
  return_as_list = FALSE,
  check_prediction_domain = "train"
)

Source

irp_holocellulose_2(), irp_klason_lignin_2()

Teickner and Knorr (2022) .

Arguments

x

An object of class ir with transmission mid infrared spectra. Some tests are applied to check if the supplied spectra match the spectra used to fit the models (the spectral range is checked). The spectral resolution of the original spectral data should not be smaller than 4 cm\(^{-1}\) and it is not checked if this assumption is met. For the following models, x has special meaning:

irp_microbial_nitrogen_content_1()

Here, x is a set of litter spectra after decomposition (and y is a set of litter spectra before decomposition). See the Details section.

...

Additional arguments passed to rstanarm::posterior_predict.stanreg() (irp_eac_1(),irp_eac_2()).

do_summary

A logical value indicating if the predicted values should be returned in a summarized version (TRUE) or not (FALSE).

• If do_summary = FALSE, a list column is returned and each element of the list column is a numeric vector, or an rvar object is returned as column in x, depending on the value of return_as_list. In both cases, the column contains draws from the posterior predictive distribution.

• If do_summary = TRUE, each element is a quantities::quantities() object with value and error summarized from posterior draws via summary_function_mean and summary_function_sd

summary_function_mean

A function used to summarize the predicted values (average).

summary_function_sd

A function used to summarize the predicted values (spread).

return_as_list

Logical value. If set to TRUE, the result will be returned as list of draws, otherwise the result will be returned as rvar object.

check_prediction_domain

A character value indicating if and how it should be checked whether the spectra in x are within the prediction domain of the model. One of:

"train"

It is checked whether the spectra in x are within the prediction domain formed by the training data for the model.

"test"

It is checked whether the spectra in x are within the prediction domain formed by the testing data for the model.

"none"

It is not checked whether the spectra in x are within the prediction domain for the model.

Details

The models use the models of the same name in the 'irpeatmodels' package. The 'irpeatmodels' package provides information on the models.

References

Teickner H, Knorr K (2022). “Improving Models to Predict Holocellulose and Klason Lignin Contents for Peat Soil Organic Matter with Mid-Infrared Spectra.” SOIL, 8(2), 699–715. doi:10.5194/soil-8-699-2022 .

See also

irp_predict()

Examples

library(ir)

x <- ir::ir_sample_data[1, ]

## make predictions

# holocellulose content
x <- irpeat::irp_holocellulose_content_2(
  x,
  do_summary = TRUE,
  check_prediction_domain = "train"
)

# Klason lignin content
x <- irpeat::irp_klason_lignin_content_2(
  x,
  do_summary = TRUE,
  check_prediction_domain = "train"
)