acf_max: Maximum ACF, PACF and CCF

In AkselA/R-ymse: Ymse (Various)

Description

Find lag that maximizes correlation

Usage

acf_max(x, ..., plot = FALSE, show = plot, ci = 0.95, ma.ci = TRUE,
  max.type = c("pos", "neg", "abs"), most.signif = FALSE)

pacf_max(x, ..., plot = FALSE, show = plot, ci = 0.95,
  max.type = c("pos", "neg", "abs"))

ccf_max(x, y, ..., plot = FALSE, show = plot, ci = 0.95,
  max.type = c("pos", "neg", "abs"))

Arguments

x, y	univariate numeric vector or time series
...	further arguments passed to acf, pacf, ccf
plot	logical; return a plot
show	indicate on the plot the maximum correlation
ci	confidence interval used, by default 95%
ma.ci	should the confidence limits assume an MA input (TRUE, the default), or white noise as is default for plot.acf?
max.type	what maximum should be returned, the positive (default), negative, or absolute maximum?
most.signif	should the most significant correlation be returned. Only applicable if ma.ci=TRUE

Examples

x <- c(5, 5, 3, 6, 3, 6, 9, 6, 3, 1, 3, 2, 8, 9, 4, 3, 6, 6,
       6, 7, 5, 2, 5, 1, 5, 5, 0, 3, 7, 3, 6, 6, 2, 2, 6, 5)
y <- c(8, 9, 7, 5, 3, 5, 6, 9, 6, 3, 4, 5, 9, 7, 8, 5, 5, 7, 
       4, 7, 7, 2, 5, 6, 5, 7, 5, 3, 5, 6, 7, 0, 5, 3, 8, 4)

acf_max(x, plot=TRUE, max.type="abs")
acf_max(x, max.type="neg")
acf_max(x, max.type="neg", most.signif=TRUE)

pacf_max(x, plot=TRUE)
pacf_max(x, max.type="abs")

ccf_max(x, y, plot=TRUE)
ccf_max(x, y, max.type="neg")

# Same plot
plot(acf(x, plot=FALSE), ci.type="ma")
acf_max(x, plot=TRUE)


acf_max(x, ci=0.99, plot=TRUE)
ccf_max(x, y, ci=0, max.type="pos", plot=TRUE)

AkselA/R-ymse documentation built on March 21, 2020, 9:52 a.m.