| Title: | ECOD025 - Problem Set 1 about VAR, FA-VAR and DFM models. Install this from the R-Universe. |
|---|---|
| Description: | Using armadillo4r for a straightforward implementation of an EM algorithm and the Kalman filter for DFM. |
| Authors: | Mauricio Vargas Sepulveda [aut, cre] (ORCID: <https://orcid.org/0000-0003-1017-7574>) |
| Maintainer: | Mauricio Vargas Sepulveda <[email protected]> |
| License: | Apache License (>= 2) |
| Version: | 0.1 |
| Built: | 2026-05-28 14:59:07 UTC |
| Source: | https://github.com/pachadotdev/ecod025 |
Estimates a Dynamic Factor Model using PCA initialization and EM algorithm with Kalman filter.
dfm_model(x, n_factors, p, max_iter, tol, forecast_h)dfm_model(x, n_factors, p, max_iter, tol, forecast_h)
x |
Matrix of observed variables (T x N) |
n_factors |
Number of latent factors (r < N) |
p |
VAR lag order for factor dynamics (0 for static model) |
max_iter |
Maximum EM iterations |
tol |
Convergence tolerance for log-likelihood |
forecast_h |
Forecast horizon (0 for no forecast) |
Estimates a FA-VAR model using PCA for factor extraction.
favar_model(y, n_lags, n_factors, p_y, p_f, include_const, forecast_h)favar_model(y, n_lags, n_factors, p_y, p_f, include_const, forecast_h)
y |
Time series vector (T x 1) |
n_lags |
Number of lags of y to include in X for factor extraction |
n_factors |
Number of latent factors to extract (r < n_lags + 1) |
p_y |
VAR lag order for y dynamics |
p_f |
VAR lag order for factor dynamics |
include_const |
Include constant term in VARs |
forecast_h |
Forecast horizon (0 for no forecast) |
Estimates a VAR model using OLS.
var_model(y, p, include_const, forecast_h)var_model(y, p, include_const, forecast_h)
y |
Time series vector (T x 1) |
p |
Lag order |
include_const |
Include constant term in VAR |
forecast_h |
Forecast horizon (0 for no forecast) |