Bayesian Prediction with High-order Interactions

Longhai Li, Department of Mathematics and Statistics, University of Saskatchewan

Copyright

Permission is granted for anyone to copy, use, modify, or distribute these programs and accompanying documents for any purpose, provided this copyright notice is retained and prominently displayed, and note is made of any changes made to these programs. These programs and documents are distributed without any warranty, express or implied. As the programs were written for research purposes only, they have not been tested to the degree that would be advisable in any important application. All use of these programs is entirely at the user's own risk.

If you found that this software is useful in your work, please do not hesitate to cite the software or the papers below.

Description

This R package is used in two situations. The first is to predict the next outcome based on the previous states of a discrete sequence. The second is to classify a discrete response based on a number of discrete covariates. In both situations, we use Bayesian logistic regression models that consider the high-order interactions. The time arising from using high-order interactions is reduced greatly by our compression technique that represents a group of original parameters as a single one in MCMC step. In this version, we use log-normal prior for the hyperparameters. When it is used for the second situation --- classification, we consider the full set of interaction patterns up to a specified order.

Source Packages and Documentations

BPHO_1.3-0, first released 23/06/2009, Documentation: BPHO_1.3-0.pdf, source package: BPHO_1.3-0.tar.gz
An important note to version earlier than 1.3

Because I read the whole Markov chain of 'betas' as needed in prediction into memory as a single static array, it may exceed the stack limit of your system, resulting error "C stack overflow". In unix or linux system, a quick way to solve this problem is to relax the stack limit before starting R, by issuing the following command in terminal:
ulimit -s unlimited
BPHO_1.2-5, first released 07/04/2008, Documentation: BPHO_1.2-5.pdf, source package: BPHO_1.2-5.tar.gz
BPHO_1.2-3, first released 22/02/2008, Documentation: BPHO_1.2-3.pdf, source package: BPHO_1.2-3.tar.gz

References

The methods used in this software are discussed in details in the following papers:

Li, L., and Neal, R. M. (2008), Compressing Parameters in Bayesian High-order Models with Applications to Sequence Models, Bayesian Analysis, 2008, volume 3, number 4, pp 793 - 822: abstract

Li, L. (2007), Bayesian Classification and Regression with High Dimensional Features, Ph.D. thesis, University of Toronto: abstract

Instruction of Installing an R package and Using R

Click here.

Back to Longhai Li's home page, or his software packages