Adjust for Linear Drift

Adjust a regular time series for an assumed linear drift over a known period while maintaining an assumed linear trend over the entire period.

adjust_linear(x, mask)

Arguments

x	A vector of data with missing values.
mask	A logical mask that identifies a subgroup of x to correct for drift. Values outside the subgroup are used to determine the overall linear trend. If no mask is specified, no overall trend is assumed.

Value

The vector x with corrected values for the subgroup specified by mask.

Examples

# no trend, only drift
x = rnorm(100, 10) + 0.1 * (1:100)
newx = adjust_linear(x)
if (interactive()) {
  plot(x, type = 'l')
  lines(newx, col = 'red')
}
# trend and drift
x = rnorm(100, 10) - 0.05 * (1:100)
mask = seq_along(x) %in% 20:80
x[mask] = x[mask] - 0.1 * cumsum(mask)[mask]
newx = adjust_linear(x, mask)
if (interactive()) {
  plot(x, type = 'l')
  lines(newx, col = 'red')
}