pastecs (version 1.1-1)

deccensus: Time decomposition using the CENSUS II method

Description

The CENSUS II method allows to decompose a regular time series into a trend, a seasonal component and residuals, according to a multiplicative model

Usage

deccensus(x, type="multiplicative", trend=FALSE)

Arguments

x

A single regular time serie (a 'rts' object under S+ and a 'ts' object under R) with a "years" time scale (one unit = one year) and a complete number of cycles (at least 3 complete years)

type

The type of model. This is for compatibility with other decxxx() functions, but only a multiplicative model is allowed here

trend

If trend=TRUE a trend component is also calculated, otherwise, the decomposition gives only a seasonal component and residuals

Value

• a 'tsd' object

Details

The trend component contains both a general trend and long-term oscillations. The seasonal trend may vary from year to year. For a seasonal decomposition using an additive model, use decloess() instead

References

B�thoux, N., M. Etienne, F. Ibanez & J.L. Rapaire, 1980. Sp�cificit�s hydrologiques des zones littorales. Analyse chronologique par la m�thode CENSUS II et estimation des �changes oc�an-atmosph�re appliqu�s � la baie de Villefranche sur mer. Ann. Inst. Oc�anogr. Paris, 56:81-95. Fromentin, J.M. & F. Ibanez, 1994. Year to year changes in meteorological features on the French coast area during the last half-century. Examples of two biological responses. Oceanologica Acta, 17:285-296. Institut National de Statistique de Belgique, 1965. D�composition des s�ries chronologiques en leurs composantes suivant diff�rentes m�thodes. Etudes statistiques et �conom�triques. Bull. Stat. INS, 10:1449-1524. Philips, J. & R. Blomme, 1973. Analyse chronologique. Universit� Catholique de Louvain, Vander ed. 339 pp. Rosenblatt, H.M., 1968. Spectral evaluation of BLS and CENSUS revised seasonal adjustment procedures. J. Amer. Stat. Assoc., 68:472-501. Shiskin, J. & H. Eisenpress, 1957. Seasonal adjustment by electronic computer methods. J. Amer. Stat. Assoc., 52:415-449.

See Also

tsd, tseries, decaverage, decdiff, decmedian, decevf, decreg, decloess

Examples

data(releve)
# Get regulated time series with a 'years' time-scale
rel.regy <- regul(releve$Day, releve[3:8], xmin=6, n=87, units="daystoyears", frequency=24, tol=2.2, methods="linear", datemin="21/03/1989", dateformat="d/m/Y")
rel.ts <- tseries(rel.regy)
# We must have complete cycles to allow using deccensus()
start(rel.ts)
end(rel.ts)
rel.ts2 <- window(rel.ts, end=c(1992,5))
rel.dec2 <- deccensus(rel.ts2[, "Melosul"], trend=TRUE)
plot(rel.dec2, col=c(1,4,3,2))