freqdom (version 2.0.5)

filter.process: Convolute (filter) a multivariate time series using a time-domain filter

Description

This function applies a linear filter to some vector time series.

Usage

filter.process(X, A)
X %c% A

Value

A matrix. Row $t$ corresponds to $Y_t$.

Arguments

X

vector time series given in matrix form. Each row corresponds to a timepoint.

A

an object of class timedom.

Functions

• filter.process(): Multivariate convolution (filter) in the time domain

• X %c% A: Convenience operator for filter.process function

Details

Let $[X_1,\dots ,X_T{]}^{\prime }$ be a $T\times d$ matrix corresponding to a vector series $X_1,\dots ,X_T$. This time series is transformed to the series $Y_1,\dots ,Y_T$, where $$Y_t=\sum _{k=-q}^p{A}_k{X}_{t-k},t\in \{p+1,\dots ,T-q\}.$$ The index $k$ of $A_k$ is determined by the lags defined for the time domain object. When index $t-k$ falls outside the domain $\{1,\dots ,T\}$ we set $X_t=\frac{1}{T}\sum _{k=1}^T{X}_k$.

See Also

timedom