Skip to content

The Bayesian Estimation, Analysis and Regression toolbox (BEAR) is a comprehensive (Bayesian Panel) VAR toolbox for forecasting and policy analysis.

Notifications You must be signed in to change notification settings

european-central-bank/BEAR-toolbox

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

❗ We just added some major changes into the main functionality of BEAR to improve its usability. To access the previous version of the code use the legacyCode branch

View BEAR 5.2 on File Exchange

GitHub® Actions
MATLAB

The BEAR toolbox

Table of Contents

Introduction

The Bayesian Estimation, Analysis and Regression toolbox (BEAR) is a comprehensive (Bayesian Panel) VAR toolbox for forecasting and policy analysis.

BEAR is a MATLAB based toolbox which is easy for non-technical users to understand, augment and adapt. In particular, BEAR includes a user-friendly graphical interface which allows the tool to be used by country desk economists.

Furthermore, BEAR is well documented, both within the code as well as including a detailed theoretical and user’s guide. BEAR includes state-of-the art applications such as FAVARs, stochastic volatility, time-varying parameters, mixed-frequency, sign and magnitude restrictions, conditional forecasts, Bayesian forecast evaluation measures, Bayesian Panel VAR using different prior distributions (for example hierarchical priors).

BEAR is specifically developed for transparently supplying a tool for state-of-the-art research and is planned to be further developed to always be at the frontier of economic research.

Use of BEAR implies acceptance of the End User Licence Agreement (EULA) for the Use of the Software “the Bayesian Estimation, Analysis and Regression (BEAR) toolbox”.

BEAR at ECB

Structure of the repository

This repository is organized as follows. All BEAR files that need to be installed by the end-users are located inside the tbx in four separate directories:

  • app contains all the files related to the user interface
  • bear has all the core MATLAB functions and classes
  • replications contains a set of Excel and settings files to replicate previous results from the literature
  • examples contains a set of functions to create settings objects for each VAR type
  • doc contains a set of PDFs with some of the BEAR documentation

The rest of the folders contain development files related to the development of BEAR which will not be copied in non-development environments. This folders are structured as follows:

  • tests contains a set of MATLAB unit tests which are automatically run within GitHub actions every time any change is pushed to the master branch.
  • resources contains the metadata of the MATLAB project bear.prj
  • release contains the definition files that allow the user compiling BEAR into a standalone application
  • images contains all the images used within the README files of the repository

Installing BEAR

For users

This section is aimed to those users indending to run BEAR, but not interested in working or ammending the code.

From MATLAB or MATLAB Online

To install the toolbox directly from MATLAB, please go to HOME > Add Ons, search for BEAR and install the toolbox. Alternatively, follow the steps below to install it from GitHub.

From GitHub

Download the latest bear.mltbx file that you will find under the Releases section on the right.

release location in GitHub

Once this file has been downloaded, you can double click it from MATLAB to install the toolbox.

double click on bear.mltbx to install the toolbox

For developers

If there was a need to modify the BEAR code, it is recommended not to modify the installed version. Instead, the best approach is to clone the repository and open the project. To open the project, you can double click on bear.prj in the main folder or rnu the command:

>> openProject('bear.prj')

Opening the MATLAB project will shadow the installed version of BEAR as long as the project is open, once the project is shut down, the installed version will again be default version. To check which verison of BEAR is currently running, you can run:

>> which BEARmain

For non-MATLAB users

Under releases, you will find an executable with the compiled application that you can install in any Windows based computer. To run the installer, just download the .EXE file and run it.

Alternatively, if you wanted to run the full version of BEAR, you can get a trial version of MATLAB here.

Getting started

Creating a settings object

To create a settings object you can use the function

>> s = BEARsettings(<VARtype>, 'ExcelFile','data.xlsx')

This will return a settings object with different properties depending on the selected VARtype.

Running BEAR from the command line

To run BEAR, please use:

>> BEARmain(s)

where s is a BEAR settings object created with the BEARsettings function.

Running BEAR interactively

From MATLAB run the command below to open the main BEAR interface.

>> BEARapp

Examples

If you want to see an example on how to run BEAR, you can run any of the following files directly:

s = bear_settings_BVAR
s = bear_settings_PANEL
s = bear_settings_SV
s = bear_settings_TVP
s = bear_settings_MF

change your inputs accordingly and then run BEARmain(s). Alternatively, if you wanted to build your own settings files, you use any of these as a template running for example:

copyfile(fullfile(bearroot(), 'examples', 'bear_settings_BVAR.m'), pwd)
edit('bear_settings_BVAR')

Documentation

For a full BEAR documentation please visit our doc page.

Distribute BEAR

With other MATLAB users

Any MATLAB user can download the latest version of BEAR from the GitHub repository. However, if you wanted to create your own custom distribution you can package it as a MATLAB toolbox as follows:

  1. Open the tbx.prj and edit the main fields such as author, version, and description.
  2. Either click on package or run:
projectFile = 'tbx.prj';
matlab.addons.toolbox.packageToolbox(projectFile)

With non-MATLAB users

If you wanted to share BEAR with someone who is not a MATLAB user, there are several routes you can take:

  1. You can use MATLAB Compiler to share the APP as a standalone program. For this, please open the Application Compiler from the toolstrip:

app toolstrip


  1. Select as MAIN FILE the appropriate BEAR app from your set of files. For example, tbx\app\BEARapp20a.

Compile app


  1. Under Files required for your application to run add the following in addition to the automatically detected ones:
  • tbx\bear\+bear\results.xlsx
  • tbx\replications\data_AAU2009.xlsx
  • tbx\replications\data_BBE2005.xlsx
  • tbx\replications\data_BvV2018.xlsx
  • tbx\replications\data_CH2019.xlsx
  • tbx\replications\data_WGP2016.xlsx
  • tbx\default_bear_data.xlsx
  1. Click on Package.
  2. You will get a subfolder with the files that you can use to redistribute the application

For external language integration

If you wanted to share specific functionality with users of other languages, you can take a look at Compiler SDK. The process is analgous to the previous one but selecting a different target.

License

Use of BEAR implies acceptance of the End User Licence Agreement (EULA) for the Use of the Software “the Bayesian Estimation, Analysis and Regression (BEAR) toolbox”. License

About

The Bayesian Estimation, Analysis and Regression toolbox (BEAR) is a comprehensive (Bayesian Panel) VAR toolbox for forecasting and policy analysis.

Resources

Stars

Watchers

Forks

Packages

No packages published