paleodata_explained_variance: Compute explained variance of an extracted signal

In paleovar/ptboxproxytools: PTBox proxy tools for time series processing and analysis

Compute explained variance of an extracted signal

Description

Compute explained variance of an extracted signal

Usage

paleodata_explained_variance(
  xin,
  signal_type = "pca",
  signal_components = 1,
  reference_signal = stats::prcomp(xin)$x[, 1]
)

Arguments

xin	Proxytibble with proxy data in 'zoo::zoo' format, or irregular time series object ('zoo::zoo'), xin can be multivariate
signal_type	Method to extract signals; Implemented methods: 'pca' Principal component analysis 'ca' Correspondance analysis 'prc' Principal curves (princurve package) 'custom_via_rda' Uses redundancy analysis to compute the explained variance of a specified signal (e.g. explained variance of AP signal for pollen assemblage record)
signal_components	Components for which explained variance should be computed
reference_signal	Specified signal, if signal_type='custom_via_rda' is selected

Value

Explained variance (given as a fraction) of the selected components (double between 0 and 1); if multiple components are specified, the joint explained variance is returned; if input is Proxytibble, output is proxytibble where explained variance is given in the proxy_data column

See Also

prcomp (from 'stats') for principal component analysis

cca (from 'vegan') for correspondance analysis

principal_curve (from 'princurve') for principal curves

rda (from 'vegan') for redundancy analysis

Examples

# Load ice core example data
library(PTBoxProxydata)
mng <- ProxyDataManager()
icecoredata <- load_set(mng,'icecore_testset',zoo_format = 'zoo')
# Compute PCA from multivariate zoo's
icecoredata_expl_var <- paleodata_explained_variance(icecoredata, 'pca')
print(paste0("Explained variance by PC1 for EDC time series: ",icecoredata_expl_var$proxy_data[[1]]))
print(paste0("Explained variance by PC1 for EPICA DML time series: ",icecoredata_expl_var$proxy_data[[2]]))
# We can do the same by first computing the PCA and then extract explained variance of the signal via redundancy analysis
icecoredata_pca <- paleodata_signal_extraction(icecoredata, 'pca')
icecoredata_expl_var_rda <- paleodata_explained_variance(icecoredata$proxy_data[[1]], 'custom_via_rda', reference_signal = icecoredata_pca$proxy_data[[1]])
print(icecoredata_expl_var_rda)

paleovar/ptboxproxytools documentation built on June 9, 2025, 1:40 a.m.