PvarExtraFun: Extra functions for `pvar` objects.

Description

Some extra functions for pvar objects that might by useful.

Usage

PvarSubset(pv, a=1, b=length(pv$x), dname=NULL)

MergePvar(..., JoinIfPossible = TRUE, dname = NULL)

Arguments

...

pvar objects that should be merged.

JoinIfPossible

logical. If TRUE (the default) the ending points of diferent vectors pvar$x will be treated as one point in result vector. Points will be joint only if they has the same value.

dname

the name of data vector of new pvar object.

a object of pvar.

a number indicating the start of the subset.

a number indicating the end of the subset.

Value

An object of the class pvar.

Details

The function PvarSubset finds the p-variation of the subset of pv$x. The function MergePvar merges pvar objects. It does not do any recalculations, it just combine the information into one object. Therefore p-variation value form MergePvar function not necessary real p-variation. The user must use this function with careful and some knowledge: by p-variation properties you can merge p-variation straightforward if joint point remains in the partition of joint set. The extremum points are such points for sure. See examples and illustration.

Examples

Run this code

#### pvSubset ####
set.seed(12)
x = BridgeT(rnorm(1000))

pv = pvar(x, 2)	

pvSubset = PvarSubset(pv, 200, 800)
pvSmall = pvar(x[200:800], 2)
pvSubset == pvSmall


op = par(mfrow = c(2, 1))
	plot(pvSubset)
	plot(pvSmall, xlim=range(pvSmall$TimeLabel), ylim=range(pvSmall$x))
par(op)

#### MergePvar ####
set.seed(13)
x = BridgeT(rnorm(1000))
S = Split_MinMax(x)

pv = pvar(x, 2)	

pv1 = pvar(x[1:S[2]], 2)	 
pv2 = pvar(x[S[2]:length(x)], 2)	 

pvJ = MergePvar(pv1, pv2)
pvJ == pv	# since joint point is extremum, we get the correct value

op = par(mfrow = c(2, 1))
	plot(pv)
	plot(pvJ, xlim=range(pv$TimeLabel), ylim=range(pv$x))
par(op)

# on the other hand if we take other points we usually do not get correct value
pvJ = MergePvar(pvar(x[1:500], 2), pvar(x[500:length(x)], 2))
pvJ == pv