ISturbulence: Calculate In-sample turbulence

In sstoeckl/uncertaintymeasures: Functions to create uncertainty measures using turbulence

Description

Calculate IS-turbulence based on a cross-section of times-series (xts) data

Usage

ISturbulence(
  X,
  weights = NULL,
  squared = FALSE,
  norm = FALSE,
  method = NULL,
  s.k = 1,
  imp = FALSE,
  use = "complete.obs",
  GW = FALSE
)

Arguments

X	the dataset X in xts format (if not, the program will try to convert it to xts)
weights	optional, can either be a vector (in which case it will be used for each row and must match ncol(X)) or a matrix with the same dimension as X if weights=NULL (the default), equal weights will be generated based on the number of columns (CHECK no of non NA entries per row)
squared	optional, should the square root be taken of the results? (standard: TRUE which correwsponds to squared=FALSE)
norm	optional, should the result be normalized by 1/sum(weights^2) per row (makes expectation equal to 1)
method	optional, if a robust mean/covariance estimator should be used
s.k	optional, lookback window for current observations that are related to the long-term mean
imp	optional, should missing values be imputed? (standard:)
use	optional, use="pairwise.complete.obs": What method of dealing with missing values should be used (note, that variables with more than half missing obs are thrown out anyway)
GW	optional, should the Garthwaite & Koch (2016) decomposition be calculated

Value

list containing 4 elements: - turb: turbulence index - turb.grad: contribution of individual series to turb - mturb: turbulence index based on diagonal covariance matrix - cturb: correlation turbulence turb/mturb - GW: Shall the Garthwaite & Koch (2016) decomposition also be calculated?

Examples

require(xts)
data(X)
turb1 <- ISturbulence(X, s.k=1)$turb
turb2 <- ISturbulence(X,s.k=12)$turb
turb3 <- ISturbulence(X,s.k=12, method="MCD")$turb
plot(turb1, lwd=2, main="In-sample turbulence (standard)")
lines(turb2,lwd=2,col="red", main="Turbulence with 12 month lookback period",on=NA)
lines(turb3,lwd=2,col="blue", main="Turbulence with 12 month lookback period and robust MCD estimator",on=NA)
turbG <- ISturbulence(X,s.k=12)$turb.grad
turbGW <- ISturbulence(X,s.k=12, GW=TRUE)$turb.GW
tail(rowSums(turbG))
tail(rowSums(turbGW))

sstoeckl/uncertaintymeasures documentation built on Nov. 7, 2021, 5:20 p.m.