Rolling sum and standard deviation

Right-aligned windows: the value of day t covers days t - width + 1 to t, and the first width - 1 days are NA. Computed from cumulative sums, so a long history takes no time.

Usage

rolling_sum(x, width)

rolling_sd(x, width)

Arguments

x

Numeric vector without NA.

width

Number of days in each window.

Value

A numeric vector as long as x.

Examples

golemcrypto:::rolling_sum(1:5, 2)
#> [1] NA  3  5  7  9
golemcrypto:::rolling_sd(c(1, 3, 2, 5, 4), 3)
#> [1]       NA       NA 1.000000 1.527525 1.527525