growth {dembase}R Documentation

Calculate average growth rates or increments.

Description

Calculate average growth rates or increments for a DemographicArray object.

Usage

growth(
  object,
  along = NULL,
  within = NULL,
  weights,
  type = c("linear", "exponential"),
  method = c("endpoints", "lm")
)

## S4 method for signature 'Counts'
growth(
  object,
  along = NULL,
  within = NULL,
  weights,
  type = c("exponential", "linear"),
  method = c("endpoints", "lm")
)

## S4 method for signature 'Values'
growth(
  object,
  along = NULL,
  within = NULL,
  weights,
  type = c("exponential", "linear"),
  method = c("endpoints", "lm")
)

Arguments

object

Object of class DemographicArray.

along

Name or index of dimension along which growth is calculated. If along is omitted, growth looks for a dimension with dimtype "time", "age", or "cohort" (in that order).

within

Name or index of dimensions within which growth is calculated. Value "." can be used as shorthand for "all dimensions other than along". Origin-destination or parent-child pairs can be referred to using their base name. See below for examples.

weights

Object of class Counts, supplying weights to be used when aggregating dimensions not included in along or within.

type

One of "exponential" or "linear". If "exponential", growth rates are calculated; if "linear", increments are calculated. Can be abbreviated. Defaults to "exponential".

method

One of "endpoints" or "lm". See below for details.

Details

The along dimension must have dimscale "Intervals" or "Points", which means that it must have dimtype "age", "cohort", or "time". The intervals or points do not need to be evenly spaced.

If there is a dimension with dimtype "iteration" and that dimension is not explicitly included in within, then it is added to within automatically.

All dimensions of object not included in along or within are aggregated before growth rates or increments are calculated.

If a weights argument is supplied when object has class Counts, the argument is ignored. If a weights argument is not supplied when object has class Values, all elements are given equal weight.

If method is "endpoints", then growth rates or increments are calculated from the first and last points only. If method is "lm", then growth rates are calculated by fitting a linear model to the logged values, and increments are calculated by fitting a linear model to the values.

Value

If within is NULL, a numeric value; otherwise an object with the same class as object.

See Also

extrapolate

Examples


## construct data
spend <- USPersonalExpenditure
names(dimnames(spend)) <- c("category", "year")
spend <- Counts(spend)

## calculate average growth rate across all categories
growth(spend, along = "year")
growth(spend) ## defaults to using dimension with dimtype "time"

## growth within categories
growth(spend, within = "category")
growth(spend, within = ".")

## increments rather than rates
growth(spend, within = "category", type = "linear")

## two different methods for calculating
growth(spend, within = "category")
growth(spend, within = "category", method = "lm")

## use base name to specific origin and destination dimensions
a <- array(rpois(n = 45, lambda = 10),
           dim = c(3, 3, 5),
           dimnames = list(reg_orig = 1:3,
               reg_dest = 1:3, time = 2000:2004))
x <- Counts(a, dimscales = c(time = "Points"))
growth(x, within = c("reg_orig", "reg_dest"))
growth(x, within = "reg")
[Package dembase version 0.0.0.119 Index]