Maxillaria

A light-weight nested sampler. The code is based on the example given in Sivia and Skilling 2006, Data Analysis, A Bayesian Tutorial, Second Edition. I have made the interface similar to the widely used nested-sampler MultiNest (Feroz et al 2008).

This is a work in progress... The code will have many bugs. If you find one please report using issue tracking.

Author

1. Sreekumar Thaithara Balan (tbs1980@gmail.com)

Compilation

Unpack the Maxillaria files

$ tar xvzf Maxillaria.tar.gz

Make a direcotry to build the files. Let's call it build. Then change directory to build

$ cd Maxillaria
$ mkdir build
$ cd build

Now we are ready to build, type

$ cmake ../
$ make

This will create a static library called libmaxillaria.a and an example excutable unit_gauss.exe.

Examples

A unit-gaussian posterior distribution case can be found in examples direcotry. First make a direcotry called chains and the run the excutable. $ mkdir chains $ ./examples/unit_gauss.exe

The output will look like

-------------------------------------------------
	     Sree's Nested Sampler
	        Version 1.0
   Sreekumar Thaithara Balan (tbs1980@gmail.com)
-------------------------------------------------
Starting sampling from scratch...

Number of iterates    = 100
Evidence,log(Z)        = -144.73 +- 0.0688597
Dlog(Z)                = 0.25315

Number of iterates    = 200
Evidence,log(Z)        = -121.72 +- 0.0659767
Dlog(Z)                = 0.14761

Number of iterates    = 300
Evidence,log(Z)        = -110.338 +- 0.0640854
Dlog(Z)                = 0.216004

Number of iterates    = 400
Evidence,log(Z)        = -98.297 +- 0.06388
Dlog(Z)                = 0.106163

................................................
................................................
................................................

Number of iterates    = 7300
Evidence,log(Z)        = -5.15544 +- 0.0634252
Dlog(Z)                = 0.000105041

Number of iterates    = 7400
Evidence,log(Z)        = -5.14546 +- 0.0634269
Dlog(Z)                = 9.48654e-05

-------------------------------------------------
Sampling finished!
Iteraions          = 7400
Evidence:log(Z)    = -5.14546 +- 0.0634269
Information: H     = 4.022972 nats(= 5.803922 bits)
-------------------------------------------------

Output files

The output files can be found in chains direcotry.

$ ls -l chains/
-rw-r--r-- 1 sree sree 592079 Nov  9 10:03 test.extract.txt
-rw-r--r-- 1 sree sree  79002 Nov  9 10:03 test.phys_live_points.txt
-rw-r--r-- 1 sree sree 289270 Nov  9 10:03 test.post_equal_weights.txt
-rw-r--r-- 1 sree sree    676 Nov  9 10:03 test.stats.txt

The *.stats.txt contains the posterior estimates.