percentChange: Various Time Series Calculations

Description

Calculate various conversions of time series.

Usage

percentChange(obj, ...)
    # S3 method for default
percentChange(obj, base=NULL, lag=1, cumulate=FALSE, e=FALSE, ...)
    ytoypc(obj, names = paste("y to y %ch", seriesNames(obj)))
    # S3 method for default
ytoypc(obj, names = paste("y to y %ch", seriesNames(obj)))
    annualizedGrowth(obj, ...)
    # S3 method for default
annualizedGrowth(obj, lag=1, freqLagRatio=frequency(obj)/lag,
        names=paste("Annual Growth of", seriesNames(obj)), ...)

Arguments

obj

An object on which the calculation is to be done. The default method works for a time series vector or matrix (with columns corresponding to series, which are treated individually).

If e is TRUE the exponent of the series is used (after cumulating if cumulate is TRUE). e can be a logical vector with elements corresponding to columns of obj.

base

If base is provided it is treated as the first period value (that is, prior to differencing). It is prefixed to the m prior to cumulating. It should be a vector of length dim(m)[2]. (If e is TRUE then base should be log of the original data).

lag

integer indicating the number of periods relative to which the change should be calculated.

cumulate

logical indicating if the series should be cumulated before the percent change is calculated.

freqLagRatio

the ratio of obj's frequency to the number of lags.

names

gives new names to be given to the calculated series.

...

arguments passed to other methods.

Value

A time series or time series matrix.

Details

percentChange calculate the percent change relative to the data lag periods prior. If cumulate is TRUE then the data is cumulated first. cumulate can be a logical vector with elements corresponding to columns of obj.

The result is a time series of the year over year percent change. This uses percentChange with lag=frequency(obj).

The names are not applied to the new series if the global option ModSeriesNames is FALSE. This can be set with options(ModSeriesNames=FALSE). This provides a convenient mechanism to prevent changing series labels on plot axis, when the title may indicate that data is in year-to-year percent change so the axis label does not need this.

annualizedGrowth calculates the year to year percentage growth rate using 100*((obj/shift(obj, periods= -lag))^freqLagRatio - 1). The default gives the annualized one period growth. If lag is equal to the frequency of obj then the result is year-over-year growth.

Examples

Run this code

# NOT RUN {
z <- ts(matrix(100 + rnorm(200),100,2), start=c(1990,1), frequency=12)
z[z == 0] <- 1 # not to likely, but it can happen
zyypc <- ytoypc(z)
zpc <- percentChange(z)
zag <- annualizedGrowth(z)
# }