tsvr (version 1.0.2)

aggts: Aggregates a tsvreq_classic object across a set of timescales; also the constructor function for class vreq_classic_ag.

Description

All the components of a tsvreq_classic object can be aggregated across an arbitrary set of timescales, producing a new variance ratio equation - this function performs that aggregation. The function returns a vreq_classic_ag object, and is the constructor function of that class. The vreq_classic_ag class has slots com, comnull, vr, which are the same as a vreq object, but also has slot ts, which is the timescales over which aggregation was performed to get the object. The class inherits from vreq, which inherits from list.

Usage

aggts(obj, ts)

Arguments

obj

A tsvreq_classic object

ts

The timescales to aggregate over

Value

aggts returns an object of class vreq_classic_ag. Slots are:

com

the timescale-aggregated value of CVcom2

comnull

the timescale-aggregated value of CVcomip2

vr

the timescale-aggregated value of the classic variance ratio

ts

the timescales over which aggregation was performed

Details

Before aggregation is performed, the argument `ts` is intersected with the canonical Fourier timescales greater than or equal to the Nyquist timescale, and the resulting timescales are then reflected about the Nyquist timescale. This is to account for the symmetry of Fourier transforms about the Nyquist frequency. The `ts` slot of the output object shows the intersected, reflected timescales that were actually used for aggregation. See the examples.

References

Zhao et al, (In prep) Decomposition of the variance ratio illuminates timescale-specific population and community variability.

See Also

tsvreq_classic, vreq_classic_ag_methods, browseVignettes("tsvr")

Examples

# NOT RUN {
X<-matrix(runif(10*100),10,100)
h<-tsvreq_classic(X)
res1<-aggts(h,h$ts[h$ts>=4]) 
res2<-aggts(h,h$ts[h$ts>=4 | h$ts<=4/3]) 
#res1 and res2 produce the same result 
#because of Fourier symmetry around the 
#Nyquist timescale - see Details 

# }