filterCoef: Coefficients and other basic properties of filters

In GeoBosh/sarima: Simulation and Prediction with Seasonal ARIMA Models

Coefficients and other basic properties of filters

Description

Coefficients and other basic properties of filters.

Usage

filterCoef(object, convention, ...)

filterOrder(object, ...)

filterPoly(object, ...)

filterPolyCoef(object, lag_0 = TRUE, ...)

Arguments

object	object.
convention	convention for the sign.
lag_0	if FALSE, drop the coefficient of order zero.
...	further arguments for methods.

Details

Generic functions to extract basic properties of filters: filterCoef returns coefficients, filterOrder returns the order, filterPoly, returns the characteristic polynomial, filterPolyCoef gives the coefficients of the characteristic polynomial.

For further details on argument convention see filterCoef-methods.

What exactly is returned depends on the specific filter classes, see the description of the corresponding methods. For the core filters, the values are as can be expected. For "ArmaFilter", the value is a list with components "ar" and "ma" giving the requested property for the corresponding part of the filter. Similarly, for "SarimaFilter" the values are lists, maybe with additional quantities.

Value

the requested property as described in Details.

Note

The filterXXX() functions are somewhat low level and technical. They should be rarely needed in routine work. The corresponding modelXXX are more flexible.

Author(s)

Georgi N. Boshnakov

See Also

modelOrder, modelCoef, modelPoly, modelPolyCoef, for the recommended higher level alternatives for models.

filterOrder-methods, filterCoef-methods, filterPoly-methods, filterPolyCoef-methods, for more information on the methods and the arguments.

Examples

filterPoly(as(c(0.3, 0.5), "BJFilter"))   # 1 - 0.3*x - 0.5*x^2
filterPoly(as(c(0.3, 0.5), "SPFilter"))   # 1 + 0.3*x + 0.5*x^2

## now two representations of the same filter:
fi1 <- as(c(0.3, 0.5), "BJFilter")
fi2 <- as(c(-0.3, -0.5), "SPFilter")
identical(fi2, fi1) # FALSE, but
## fi1 and fi2 represent the same filter, eg. same ch. polynomials:
filterPoly(fi1)
filterPoly(fi2)
identical(filterPolyCoef(fi2), filterPolyCoef(fi1))

# same as above, using new()
fi1a <- new("BJFilter", coef = c(0.3, 0.5))
identical(fi1a, fi1) # TRUE

fi2a <- new("SPFilter", coef = c(-0.3, -0.5))
identical(fi2a, fi2) # TRUE

## conversion by as() changes the internal representation
## but represents the same filter:
identical(as(fi1, "SPFilter"), fi2) # TRUE

c(filterOrder(fi1), filterOrder(fi2))

## these give the internally stored coefficients:
filterCoef(fi1)
filterCoef(fi2)

## with argument 'convention' the result doesn't depend
## on the internal representation:
co1 <- filterCoef(fi1, convention = "SP")
co2 <- filterCoef(fi2, convention = "SP")
identical(co1, co2) # TRUE

GeoBosh/sarima documentation built on March 27, 2024, 6:31 p.m.