Last chance! 50% off unlimited learning
Sale ends in
Detrends y
by subtracting a linear trend in x
, to create
a vector that is zero for its first and last finite value.
If the second parameter (y
) is missing, then x
is
taken to be y
, and a new x
is constructed with
seq_along
. Any NA
values are left as-is.
detrend(x, y)
a vector of numerical values. If y
is not given, then
x
is taken for y
.
an optional vector
A list containing Y
, the detrended version of y
, and
the intercept a
and slope b
of the linear function of x
that is subtracted from y
to yield Y
.
A common application is to bring the end points of a time series down to zero, prior to applying a digital filter. (See examples.)
# NOT RUN {
x <- seq(0, 0.9 * pi, length.out=50)
y <- sin(x)
y[1] <- NA
y[10] <- NA
plot(x, y, ylim=c(0, 1))
d <- detrend(x, y)
points(x, d$Y, pch=20)
abline(d$a, d$b, col='blue')
abline(h=0)
points(x, d$Y + d$a + d$b * x, col='blue', pch='+')
# }
Run the code above in your browser using DataLab