moving: Moving Functions

In armstrtw/fts: R interface to tslib (a time series library in c++)

Description

apply summary functions on a moving/rolling basis

Usage

moving.mean(x, periods)
moving.sum(x, periods)
moving.max(x, periods)
moving.min(x, periods)
moving.sd(x, periods)
moving.rank(x, periods)

moving.cor(x, y, periods)
moving.cov(x, y, periods)

cor.by.row(x,y)

Arguments

x	An Fts object
y	An Fts object
periods	integer: number of periods in window

Details

apply a function that takes a vector and returns a scalar on a rolling basis to an fts object.

For cor.by.row, the indicator is not rolling, but is the result of the application of the cor function to matching rows of x and y.

asking for a window larger than the number of rows of the fts object will result in an fts of all NA w/ the same number of rows as the input

for functions that take two fts objects the date intersection is taken before the window function is applied

Value

an fts object

Author(s)

Whit Armstrong

Examples

x <- fts(index=seq(from=Sys.Date(),by="days",length.out=100),data=1:100)
y <- fts(index=seq(from=Sys.Date(),by="days",length.out=100),data=1:100)

x.mean <- moving.mean(x,20)
x.sum <- moving.sum(x,20)
x.prod <- moving.product(x,20)
x.max <- moving.max(x,20)
x.min <- moving.min(x,20)
x.sd <- moving.sd(x,20)
x.rank <- moving.rank(x,20)

## take only odd rows
## to illustrate that teh correlation and covariance
## will only be calculated for the intersection of the dates
y <- y[(1:nrow(y))%%2 == 1]

xy.cor <- moving.cor(x, y, 20)
xy.cov <- moving.cov(x, y, 20)

armstrtw/fts documentation built on May 10, 2019, 1:42 p.m.