filter.twoSided: Apply a linear filter to a time-series

In driegert/transfer3: Estimate transfer functions using non-stationary multitaper framework

Description

Assumes a two-sided filter.

Usage

filter.twoSided(x, filter)

Arguments

x	A vector containing the series to filter.
filter	A vector containing the filter coefficients. See details.

Details

Ends up being twice as fast as stats::filter(). Only 2-sided. Only convolution.

The argument, filter, should be of length 2*L + 1 (i.e., odd). The positive lag (causal) filter coefficients should be in elements (L+1):(2*L) with the negative lag (acausal) filter coefficients contained in the 1:L elements of filter. The L+1 index of filter is lag-0. This is the same convention as required by stats::filter().

driegert/transfer3 documentation built on May 3, 2019, 8:33 p.m.