zooreg: Regular zoo Series

In parsifal9/test: S3 Infrastructure for Regular and Irregular Time Series (Z's Ordered Observations)

Description

zooreg is the creator for the S3 class "zooreg" for regular "zoo" series. It inherits from "zoo" and is the analogue to ts.

Usage

zooreg(data, start = 1, end = numeric(), frequency = 1, 
  deltat = 1, ts.eps = getOption("ts.eps"), order.by = NULL,
  calendar = getOption("zoo.calendar", TRUE))

Arguments

data	a numeric vector, matrix or a factor.
start	the time of the first observation. Either a single number or a vector of two integers, which specify a natural time unit and a (1-based) number of samples into the time unit.
end	the time of the last observation, specified in the same way as start.
frequency	the number of observations per unit of time.
deltat	the fraction of the sampling period between successive observations; e.g., 1/12 for monthly data. Only one of frequency or deltat should be provided.
ts.eps	time series comparison tolerance. Frequencies are considered equal if their absolute difference is less than ts.eps.
order.by	a vector by which the observations in x are ordered. If this is specified the arguments start and end are ignored and zoo(data, order.by, frequency) is called. See zoo for more information.
calendar	logical. Should yearqtr or yearmon be used for a numeric time index with frequency 4 or 12, respectively?

Details

Strictly regular series are those whose time points are equally spaced. Weakly regular series are strictly regular time series in which some of the points may have been removed but still have the original underlying frequency associated with them. "zooreg" is a subclass of "zoo" that is used to represent both weakly and strictly regular series. Internally, it is the same as "zoo" except it also has a "frequency" attribute. Its index class is more restricted than "zoo". The index: 1. must be numeric or a class which can be coerced via as.numeric (such as yearmon, yearqtr, Date, POSIXct, tis, xts, etc.). 2. when converted to numeric must be expressible as multiples of 1/frequency. 3. group generic functions Ops should be defined, i.e., adding/subtracting a numeric to/from the index class should produce the correct value of the index class again.

zooreg is the zoo analogue to ts. The arguments are almost identical, only in the case where order.by is specified, zoo is called with zoo(data, order.by, frequency). It creates a regular series of class "zooreg" which inherits from "zoo". It is essentially a "zoo" series with an additional "frequency" attribute. In the creation of "zooreg" objects (via zoo, zooreg, or coercion functions) it is always check whether the index specified complies with the frequency specified.

The class "zooreg" offers two advantages over code "ts": 1. The index does not have to be plain numeric (although that is the default), it just must be coercible to numeric, thus printing and plotting can be customized. 2. This class can not only represent strictly regular series, but also series with an underlying regularity, i.e., where some observations from a regular grid are omitted.

Hence, "zooreg" is a bridge between "ts" and "zoo" and can be employed to coerce back and forth between the two classes. The coercion function as.zoo.ts returns therefore an object of class "zooreg" inheriting from "zoo". Coercion between "zooreg" and "zoo" is also available and drops or tries to add a frequency respectively.

For checking whether a series is strictly regular or does have an underlying regularity the generic function is.regular can be used.

Methods to standard generics for regular series such as frequency, deltat and cycle are available for both "zooreg" and "zoo" objects. In the latter case, it is checked first (in a data-driven way) whether the series is in fact regular or not.

as.zooreg.tis has a class argument whose value represents the class of the index of the zooreg object into which the tis object is converted. The default value is "ti". Note that the frequency of the zooreg object will not necessarily be the same as the frequency of the tis object that it is converted from.

Value

An object of class "zooreg" which inherits from "zoo". It is essentially a "zoo" series with a "frequency" attribute.

See Also

zoo, is.regular

Examples

## equivalent specifications of a quarterly series
## starting in the second quarter of 1959.
zooreg(1:10, frequency = 4, start = c(1959, 2))
as.zoo(ts(1:10, frequency = 4, start = c(1959, 2)))
zoo(1:10, seq(1959.25, 1961.5, by = 0.25), frequency = 4)

## use yearqtr class for indexing the same series
z <- zoo(1:10, yearqtr(seq(1959.25, 1961.5, by = 0.25)), frequency = 4)
z
z[-(3:4)]

## create a regular series with a "Date" index
zooreg(1:5, start = as.Date("2000-01-01"))
## or with "yearmon" index
zooreg(1:5, end = yearmon(2000))

## lag and diff (as diff is defined in terms of lag)
## act differently on zoo and zooreg objects!
## lag.zoo moves a point to the adjacent time whereas
## lag.zooreg moves a point by deltat
x <- c(1, 2, 3, 6)
zz <- zoo(x, x)
zr <- as.zooreg(zz)
lag(zz, k = -1)
lag(zr, k = -1)
diff(zz)
diff(zr)

## lag.zooreg wihtout and with na.pad
lag(zr, k = -1)
lag(zr, k = -1, na.pad = TRUE)

## standard methods available for regular series
frequency(z)
deltat(z)
cycle(z)
cycle(z[-(3:4)])

zz  <-  zoo(1:6, as.Date(c("1960-01-29", "1960-02-29", "1960-03-31",
  "1960-04-29", "1960-05-31", "1960-06-30")))
# this converts zz to "zooreg" and then to "ts" expanding it to a daily
# series which is 154 elements long, most with NAs.
## Not run: 
length(as.ts(zz)) # 154

## End(Not run)
# probably a monthly "ts" series rather than a daily one was wanted.
# This variation of the last line gives a result only 6 elements long.
length(as.ts(aggregate(zz, as.yearmon, c))) # 6

zzr <- as.zooreg(zz)

dd <- as.Date(c("2000-01-01", "2000-02-01", "2000-03-01", "2000-04-01"))
zrd <- as.zooreg(zoo(1:4, dd))

parsifal9/test documentation built on Dec. 31, 2020, 1:14 a.m.