RecordTest (version 2.2.0)

records: Record Values and Record Times

Description

This function identifies (and plots if argument plot = TRUE) the record values (\(R_i\)), and the record times (\(L_i\)) in a vector, for all upper and lower records in forward and backward directions.

Usage

records(
  X,
  plot = TRUE,
  direction = c("forward", "backward", "both"),
  variable,
  type = c("lines", "points"),
  col = c(T = "black", U = "salmon", L = "skyblue", O = "black"),
  alpha = c(T = 1, U = 1, L = 1, O = 1),
  shape = c(F = 19, B = 4, O = 19),
  linetype = c(F = 1, B = 2)
)

Value

If plot = TRUE a ggplot object, otherwise a list with four data frames where the first column are the record times, the second the record values and, if variable is not null, the third column are their values at the record times, respectively for upper and lower records in forward and backward series.

Arguments

X

A numeric vector.

plot

Logical. If TRUE (the default) the records are plotted.

direction

A character string indicating the type of record to show in the plot if plot == TRUE: "forward", "backward" or "both" (see Details).

variable

Optional. A vector, containing other variable related to X and measured at the same times. Only used if plot = FALSE.

type

Character string indicating if X is shown with "lines" or "points".

col, alpha

Character and numeric vectors of length four, respectively. These arguments represent respectively the colour and transparency of the points or lines: trivial record, upper records, lower records and observations respectively. Vector names in the default are only indicative.

shape

If type == "points". Integer vector of length 3 indicating the shape of the points for forward records, backward records and observations. Vector names in the default are only indicative.

linetype

Integer vector of length 2 indicating the line type of the step functions in the forward and backward records, respectively. Vector names in the default are only indicative.

Author

Jorge Castillo-Mateo

Details

Customarily, the records in a time series \((X_t)\) observed in \(T\) instances \(t = 1, 2, \ldots, T\) can be obtained using chronological order. Besides, we could also compute the records in similar sequences of random variables if we consider reversed chronological order starting from the last observation, i.e., \(t' = T, \ldots, 2, 1\). The analysis of series with reversed order is customarily referred to as backward, as opposed to a forward analysis.

See Also

I.record, series_double, series_rev, series_split, series_uncor, series_untie

Examples

Y <- c(5, 7, 3, 6, 19, 2, 20)
records(Y, plot = FALSE, variable = seq_along(Y))

# Show the whole series and its upper and lower records
records(TX_Zaragoza$TX)
# Compute tables for the whole series
TxZ.record <- records(TX_Zaragoza$TX, plot = FALSE, variable = TX_Zaragoza$DATE)
TxZ.record
names(TxZ.record)
# To show the Forward Upper records
TxZ.record[[1]]
plot(TxZ.record[[1]]$Times, TxZ.record[[1]]$Values)

# Annual maximum daily maximum temperatures
TxZ <- apply(series_split(TX_Zaragoza$TX), 1, max)
# Plot for the records in forward and backward directions
records(TxZ, direction = "both")
# Compute tables for the annual maximum 
records(TxZ, plot = FALSE, variable = 1951:2020)