Skip to contents
dfms-package dfms
Dynamic Factor Models

Information Criteria

Choose the number of factors and the lag-order of the factor VAR.

ICr() print(<ICr>) plot(<ICr>) screeplot(<ICr>)
Information Criteria to Determine the Number of Factors (r)

Fit a Dynamic Factor Model

DFM estimation via the EM algorithm and PCA, and various methods inspect the model and extract results.

DFM()
Estimate a Dynamic Factor Model
print(<dfm>) coef(<dfm>) logLik(<dfm>) summary(<dfm>) print(<dfm_summary>)
DFM Summary Methods
plot(<dfm>) screeplot(<dfm>)
Plot DFM
as.data.frame(<dfm>)
Extract Factor Estimates in a Data Frame
residuals(<dfm>) fitted(<dfm>)
DFM Residuals and Fitted Values

Forecasting

Forecast both the factors and the data, and methods to visualize forecasts and extract results.

News Decomposition

Decompose forecast revisions into news contributions follwing Banbura and Modugno (2014).

Fast Stationary Kalman Filtering and Smoothing

Optimized Armadillo C++ implementations of the stationary Kalman Filter and Smoother.

SKF()
(Fast) Stationary Kalman Filter
FIS()
(Fast) Fixed-Interval Smoother (Kalman Smoother)
SKFS()
(Fast) Stationary Kalman Filter and Smoother

Helper Functions

Fast VAR, matrix inverses, imputation/removal of missing values in multivariate time series, and convergence check for EM algorithm.

.VAR()
(Fast) Barebones Vector-Autoregression
tsnarmimp()
Remove and Impute Missing Values in a Multivariate Time Series
ainv() apinv()
Armadillo's Inverse Functions
em_converged()
Convergence Test for EM-Algorithm

Data

Euro area macroeconomic data from Banbura and Modugno (2014), and 3 DFM specifications considered in their paper.

BM14_Models BM14_M BM14_Q
Euro Area Macroeconomic Data from Banbura and Modugno 2014