aba_combine_strata {lidaRtRee}

Combines a list of ABA models into a single ABA model object

Description

Combines a list of models (obtained with aba_build_model) into a single object. Typically used to merge stratum-specific models into one object. Validation statistics are computed for the combined strata, making it easier to compare prediction performance with an unstratified model.

Usage

aba_combine_strata(model.list, plotsId = NULL)

Arguments

model.list

list. stratum-specific models returned by aba_build_model

plotsId

vector. "plotsId" for ordering row names in the "values" element of the output list

Value

a list with three elements

• model: a list of regression models corresponding to each stratum (output from lm),

• stats: model statistics of each stratum-specific model (as in aba_build_model) plus one line corresponding to statistics for all strata (COMBINED)

• values: data.frame with observed and values predicted in cross-validation, and information on which stratum it belongs to.

See Also

aba_build_model for calibrated ABA model, aba_plot for plotting model cross-validation results.

Examples

# load Quatre Montagnes dataset
data(quatre_montagnes)
# initialize list of models
model_aba_stratified <- list()
# calibrate basal area prediction model for each stratum
for (i in levels(quatre_montagnes$stratum))
{
  subsample <- which(quatre_montagnes$stratum == i)
  model_aba_stratified[[i]] <-
    aba_build_model(quatre_montagnes[subsample, "G_m2_ha"],
      quatre_montagnes[subsample, 9:76],
      transform = "boxcox", nmax = 4,
      xy = quatre_montagnes[subsample, c("X", "Y")]
    )
}
# combine models in single object
model_aba_stratified <- aba_combine_strata(
  model_aba_stratified,
  quatre_montagnes$plotId
)
# display content of output list
model_aba_stratified$model
model_aba_stratified$stats
summary(model_aba_stratified$values)

# plot field values VS predictions in cross-validation
aba_plot(model_aba_stratified)