Bayesian Data Analysis

   Version 1.1, November 2023

   Download: BDA_Matlab_1.1.zip 

   Download: BDA_Papers_1.1.zip 

  Matlab Code for

• Autoregressive Time Series Modelling

• Linear Dynamical Systems (LDA)

• Bayesian Univariate Linear Regression, ANOVAs, t-tests

• Mixed Effects Models

• Mixtures of Factor Analysers (MFA)

• Mixtures of Canonical Correlate Analysers (MCCA)

• Gaussian Mixture Models

• Multivariate Linear Models

• Neighbourhood Component Analysis

• Principal Component Analysis

• Robust General Linear Models

• Softmax Regression

• Spectral Estimation

Some of this software was previously available in SPM12 

It also includes more recent edits, developments and contributions from third parties with wrappers to their original code. Run bda_setpath.m and then you're off ! Read Contents.m in relevant subdirectory and run demos.

Third party contributions include:

LDS and MFA by Matt Beal, UCL, 2003

MCCA code by J Viinikanoja, University of Helsinki, 2010. The original

MCCA code also uses Tom Minka's Lightspeed Toolbox, Microsoft, 2017.

The included SPM12 functions are written by various authors at UCL including Karl Friston, Guillaume Flandin, Andrew Holmes and Jean-Baptiste Poline.

PREVIOUS VERSION:

   Version 1.0, August 12 2020

   Download: BDA_Matlab_1.0.zip 

   Download: BDA Papers_1.0.zip 

Game Theory Equilibria

Equilibrium Density approaches have been proposed as normative models for how people behave in two player games. This toolbox provides matlab functions for computing equilbrium densities based on the concepts of Mixed Nash Equilibria, Quantal Response Equilibria and Gibbs Equilibria. Functions are provided for arbitrary pay-off matrices (utlities/rewards) and include examples for Stag Hunts, Inspection and Negotiation Games, showing how the equilibria depend on decision noise and parameters of the pay-off matrices.

Matlab code can be downloaded here:  equilibria_1.0.zip  and an accompanying tech report here: equilibrium.pdf