Did you know ... | Search Documentation: |
Packs (add-ons) for SWI-Prolog |
Title: | Bayesian inference of model structure. |
---|---|
Rating: | Not rated. Create the first rating! |
Latest version: | 3.0 |
SHA1 sum: | a2ab624036aedcba3b998868f09d532ef4cd92c7 |
Author: | James Cussens http://www.cs.york.ac.uk/~jc |
Nicos Angelopoulos http://stoics.org.uk/~nicos | |
Maintainer: | Nicos Angelopoulos http://stoics.org.uk/~nicos |
Packager: | Nicos Angelopoulos http://stoics.org.uk/~nicos |
Home page: | http://stoics.org.uk/~nicos/sware/bims |
Download URL: | http://stoics.org.uk/~nicos/sware/packs/bims/bims-*.tgz |
Requires: | lib |
No reviews. Create the first review!.
Version | SHA1 | #Downloads | URL |
---|---|---|---|
0.6 | 226a3c3f955f5101cf8f3f75ab326b6312c183d0 | 8 | http://stoics.org.uk/~nicos/sware/packs/bims/bims-0.6.tgz |
1.1 | 54d2565187150656160b1dcaed6dad4b2339925c | 9 | http://stoics.org.uk/~nicos/sware/packs/bims/bims-1.1.tgz |
2.0 | 746110ff7e22f090d35254a61bd333e1777ab6a8 | 3 | http://stoics.org.uk/~nicos/sware/packs/bims/bims-2.0.tgz |
2.1 | 9332c27bf7c7287b505be5cbcda568d4641ee871 | 2 | http://www.stoics.org.uk/~nicos/sware/packs/bims/bims-2.1.tgz |
2.2 | 52a1e0efb5e91208b8b38675dc3154b04edd43fb | 14 | http://stoics.org.uk/~nicos/sware/packs/bims/bims-2.2.tgz |
2.3 | 03141d2e2d24ec6b0239462714150017ad9ef10e | 30 | http://stoics.org.uk/~nicos/sware/packs/bims/bims-2.3.tgz |
2.4 | e3afa0bd6402144354ecb12dc819598dd9693fea | 1 | https://stoics.org.uk/~nicos/sware/packs/bims/bims-2.4.tgz |
2.5 | 0a4c23a03e07fcf096d03130dc81c78f9d446349 | 22 | http://stoics.org.uk/~nicos/sware/packs/bims/bims-2.5.tgz |
3.0 | a2ab624036aedcba3b998868f09d532ef4cd92c7 | 9 | https://stoics.org.uk/~nicos/sware/packs/bims/bims-3.0.tgz |
Bims (Bayesian inference over model structures) implements MCMC learning over statistical models defined in the Dlp (Distributional logic programming) probabilistic language.
Bims is released under the MIT licence.
Currently there are 2 model spaces supported:
Additional model spaces can be easily implemented by defining new likelihood plug-ins and programming appropriate priors.
?- bims( [] ). ?- bims( [data(carts),models(carts),likelihood(carts)] ).
The above are two equivalent ways to run the Carts example provided.
This runs 3 chains each of length 100 on the default Carts data using the default likelihood. The default dataset is the breast cancer Winsconsin (BCW) data from the machine learning repository. There are 2 categories, 9 variables and 683 data points in this dataset. You can view the data with
?- edit( pack(bims/data/carts) ).
The default likelihood is an implementation of the classification likelihood function presented in: H Chipman, E George, and R McCulloch. Bayesian CART model search (with discussion). J. of the American Statistical Association, 93:935â960, 1998.
?- bims( [models(bns)] ). ?- bims( [data(bns),models(bns),likelihood(bns)] ).
The above are two equivalent ways to run the Bns example provided.
This runs 3 chains each of length 100 on the default bns data using default likelihood. The dataset is a sampled dataset from the ASIA network and it comprises of 8 variables and 2295 datapoints. You can view the data with
?- edit( pack(bims/data/bns) ).
The default BN likelihood is an instance of the BDeu metric for scoring BN structures.
W. L. Buntine. Theory refinement of Bayesian networks. In Bruce DâAmbrosio, Philippe Smets, and Piero Bonissone, editors, Proceedings of the Seventh Annual Conference on Uncertainty in Artificial Intelligence (UAIâ1991), pages 52â60, 1991
David Heckerman, Dan Geiger, and David M. Chickering. Learning Bayesian networks: The combination of knowledge and statistical data. Machine Learning, 20(3):197â243, 1995.
An easy way to run Bims on your data is to create a new directory and within that sub-directory data/. Copy your data there and pass options data/1 to the basename of the data file.
For example, assuming your data are in data/bcw.pl
?- bims( [models(carts),data(bcw)] ).
To learn new statistical models create a subdirectory within models/ with new model name and place within there directories lklhoods and dlps containing code implementing the likelihood and prior progrms respectively.
See models/carts and models/bns for examples.
Pack contains 169 files holding a total of 599K bytes.