Calculate Sample Specific Errors

Calculate Sample Specific Errors

sse.sample(
  modern_taxa,
  modern_climate,
  fossil_taxa,
  trainfun,
  predictfun,
  nboot,
  nPLS,
  nsig,
  usefx = FALSE,
  fx_method = "bin",
  bin = NA,
  cpus = 4,
  seed = NULL,
  test_mode = FALSE,
  test_it = 5
)

Arguments

modern_taxa

The modern taxa abundance data, each row represents a sampling site, each column represents a taxon.

modern_climate

The modern climate value at each sampling site

fossil_taxa

Fossil taxa abundance data to reconstruct past climates, each row represents a site to be reconstructed, each column represents a taxon.

trainfun

Training function you want to use, either WAPLS.w or TWAPLS.w.

predictfun

Predict function you want to use: if trainfun is WAPLS.w, then this should be WAPLS.predict.w; if trainfun is TWAPLS.w, then this should be TWAPLS.predict.w.

nboot

The number of bootstrap cycles you want to use.

nPLS

The number of components to be extracted.

nsig

The significant number of components to use to reconstruct past climates, this can be obtained from the cross-validation results.

usefx

Boolean flag on whether or not use fx correction.

fx_method

Binned or p-spline smoothed fx correction: if usefx = FALSE, this should be NA; otherwise, fx function will be used when choosing "bin"; fx_pspline function will be used when choosing "pspline".

bin

Binwidth to get fx, needed for both binned and p-splined method. if usefx = FALSE, this should be NA;

cpus

Number of CPUs for simultaneous iterations to execute, check parallel::detectCores() for available CPUs on your machine.

seed

Seed for reproducibility.

test_mode

Boolean flag to execute the function with a limited number of iterations, test_it, for testing purposes only.

test_it

Number of iterations to use in the test mode.

Value

The bootstrapped standard error for each site.

See also

fx, TWAPLS.w, TWAPLS.predict.w, WAPLS.w, and WAPLS.predict.w

Examples

if (FALSE) {
# Load modern pollen data
modern_pollen <- read.csv("/path/to/modern_pollen.csv")

# Extract taxa
taxaColMin <- which(colnames(modern_pollen) == "taxa0")
taxaColMax <- which(colnames(modern_pollen) == "taxaN")
taxa <- modern_pollen[, taxaColMin:taxaColMax]

# Load reconstruction data
Holocene <- read.csv("/path/to/Holocene.csv")
taxaColMin <- which(colnames(Holocene) == "taxa0")
taxaColMax <- which(colnames(Holocene) == "taxaN")
core <- Holocene[, taxaColMin:taxaColMax]

## SSE
nboot <- 5 # Recommended 1000
nsig <- 3 # This should be got from the random t-test of the cross validation

sse_tf_Tmin2 <- fxTWAPLS::sse.sample(
  modern_taxa = taxa,
  modern_climate = modern_pollen$Tmin,
  fossil_taxa = core,
  trainfun = fxTWAPLS::TWAPLS.w2,
  predictfun = fxTWAPLS::TWAPLS.predict.w,
  nboot = nboot,
  nPLS = 5,
  nsig = nsig,
  usefx = TRUE,
  fx_method = "bin",
  bin = 0.02,
  cpus = 2,
  seed = 1
)

# Run with progress bar
`%>%` <- magrittr::`%>%`
sse_tf_Tmin2 <- fxTWAPLS::sse.sample(
  modern_taxa = taxa,
  modern_climate = modern_pollen$Tmin,
  fossil_taxa = core,
  trainfun = fxTWAPLS::TWAPLS.w2,
  predictfun = fxTWAPLS::TWAPLS.predict.w,
  nboot = nboot,
  nPLS = 5,
  nsig = nsig,
  usefx = TRUE,
  fx_method = "bin",
  bin = 0.02,
  cpus = 2,
  seed = 1
) %>% fxTWAPLS::pb()
}