#Filzbach

Fit complex models to heterogeneous data: Bayesian and likelihood analysis made easy

Filzbach is a flexible, fast, robust, parameter estimation engine that allows you to parameterize arbitrary, non-linear models, of the kind that are necessary in biological sciences, against multiple, heterogeneous data sets. Filzbach allows for Bayesian parameter estimation, maximum likelihood analysis, priors, latents, hierarchies, error propagation and model selection, from just a few lines of code.

Home web site: http://research.microsoft.com/filzbach

Filzbach User Guide

#Build Filzbach library

##Windows

To build static library filzbach.lib in "Debug" configuration from command prompt:

cd filzbach
msbuild

To build a dynamically linked library filzbach.dll in "Release" configuration:

cd filzbach
msbuild /p:Configuration=Release_DLL

To build either a static or a dynamic library with support for OpenMP set the CL environment variable which affects Visual C++ compiler options, e.g.

set CL=/openmp
cd filzbach
msbuild /p:Configuration=Release_DLL

##Unix

Build a library libfilzbach.a from a shell prompt:

cd filzbach
make

#Run examples

The repository contains several sample programs that illustrate how to apply Filzbach to different kinds of models.

Id	Name	Desciption
1	normal	How to learn parameters of a normal distribution
2	normal_priors	How to set up priors
3	mixednormal	Learn the location and width of several peaks
4	poisson	Learn parameters of a Poisson distribution
5	poisson_multispp	Learn parameters of several Poisson distributions at once with hierarchical modelling
7	lr	Linear regression
8	logistic_regression	Logistic regression
9	stepfunction	Step function
11	plantgrowth	Infer parameters of a simple plant growth model from observations
12	seedrain	Seed
13	SDM	Species distribution modelling
16	ProteinInteraction	Simple model of a protein interaction network
17	TreeMort	A model of tree mortality

On Windows, the easiest is to start Microsoft Visual Studio, open filzbach.sln, uncomment one of the lines in examples/examples.h header file and run the examples project.

You can also compile and run from command prompt. You then first set set up the CL environment variable with an example Id. E.g., to run species distribution modelling sample (Id=13) do the following:

set CL=/DMODEL#13
msbuild /t:Clean
msbuild
Debug\examples

On Unix systems from the examples directory use make to build an individual example using its name or build all the examples with one command:

cd examples
make examples