Description Usage Arguments Value Examples
Function that applies the Kalman filter to a univariate stationary zero-mean ARMA process. To get the Kalman filter to work the process should be written as a dynamic linear model. The ARMA process written as a dynamic linear model has the form, where y_t = F\bm{x}_t is the observation equation and \bm{x}_t = G\bm{x}_{t-1} + H w_t is the state equation, and Q = \mbox{Var}(H w_t).
1 | kalman_filter_arma(ts, F, G, Q, m0, C0)
|
ts |
A univariate time series with zero mean |
F |
The coefficient matrix in the observation equation, as shown above. |
G |
The matrix in the state equation as shown above. |
Q |
The variance matrix of the state equation. |
m0 |
The initial value of \bm{x}_t. |
C0 |
The initial value of the state variance. |
The return is a list of the innovations, standardised residuals, and predicted values.
innovations |
The innovations. |
sd |
the standardised residuals. |
y_predicted |
the predicted values. |
1 | kalman_filter_arma(ts = data, F=F, G=G, Q=Q, m0=m0, C0=C0)
|
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.