convolve_at: Convolution at Arbitrary Indices

In kuwisdelu/matter: Out-of-core statistical computing and signal processing

Convolution at Arbitrary Indices

Description

Convolve a signal with weights at arbitrary indices.

Usage

convolve_at(x, index, weights, margin = NULL, na.rm = FALSE)

Arguments

x	A numeric vector or matrix.
index	A list or matrix of numeric vectors giving the indices to convolve. If index is a list, then it must have the same length as x and element lengths that match weights. If x is a matrix, then its rows must correspond to x and its columns must correspond to weights.
weights	A list giving the weights of the kernels to convolve for each element of x. Lengths must match index.
margin	If x is a matrix, then the margin to convolve over.
na.rm	Should missing values (including NaN) be removed?

Details

This is essentially just a weighted sum defined by x[i] = sum(weights[[i]] * x[index[[i]]]).

Value

A numeric vector the same length as x with the smoothed result.

Author(s)

Kylie A. Bemis

Examples

set.seed(1)
t <- seq(from=0, to=6 * pi, length.out=5000)
y <- sin(t) + 0.6 * sin(2.6 * t)
x <- y + runif(length(y))

i <- roll(seq_along(x), width=15)
wt <- dnorm((-7):7, sd=7/2)
wt <- wt / sum(wt)

xs <- convolve_at(x, i, wt)

plot(x, type="l")
lines(xs, col="red")

kuwisdelu/matter documentation built on Dec. 8, 2024, 8:09 p.m.