Welcome to the FAVAR (Factor Augmented Vector Auto-Regression) R implementation repository! This repository provides a comprehensive implementation of a FAVAR model for forecast prediction tasks in R.

• Introduction
• Installation
• Usage
  • Function Parameters
  • Example
• Dependencies
• Contributing
• License

FAVAR (Factor Augmented Vector Auto-Regression) is an advanced econometric model that extends the standard VAR (Vector Auto-Regression) model by incorporating factors extracted from a large dataset. This approach can improve forecast accuracy by leveraging a broader information set.

This implementation allows you to fit a FAVAR model and make forecasts using your own dataset in a seamless manner.

To get started, you need to clone the repository and install the required R packages:

git clone https://github.com/PetrGarm/FAVAR.git
cd FAVAR/Code/my experiments/FAVAR_CV.R

Next, ensure you have the necessary R packages installed:

install.packages(c("tidyverse", "forecast", "vars", "lmtest"))

The main function provided by this repository is fore_FAVAR, which takes in your dataset and parameters and returns the forecast results.

• X: A matrix or data frame containing the predictors.
• Y: A matrix or data frame containing the dependent variable(s).
• K: The number of factors to extract from X.
• y_name: The name of the dependent variable in Y to be forecasted.
• h: The forecast horizon.
• y: (Optional and must not be changed probably) A time series object of the dependent variable. Defaults to Y[,y_name].
• use_VAR: (Optional) Logical flag to use standard VAR without factors if TRUE. Defaults to FALSE.

# Sample usage of the fore_FAVAR function
source("FAVAR_CV.R")


# Define your data
X <- ...  # Your predictors matrix
Y <- ...  # Your dependent variables matrix
K <- 3    # Number of factors to extract
y_name <- "target_variable"  # The name of the dependent variable to forecast
h <- 12   # Forecast horizon

# Run FAVAR
forecast_result <- fore_FAVAR(X, Y, K, y_name, h)

# Display forecast results
print(forecast_result)

This repository is organized into the following directories:

• Code: Contains the R scripts and code implementations
• Diploma: Includes any diploma-related documents and resources
• Literature: Stores research papers, articles, and other relevant literature
• Project Proposals: Contains project proposal documents and related resources
• plots: Stores plot images generated from the analysis

The following R packages are required to run the FAVAR implementation:

• tidyverse
• forecast
• vars
• lmtest

Ensure they are installed and loaded into your R environment before running the fore_FAVAR function.

Contributions to this repository are welcome. If you have any improvements, bug fixes, or new features, feel free to open a pull request or issue.

This project is licensed under the MIT License. See the LICENSE file for more details.