Description
Control of MCMC annealing parameters needed in logreg.
Usage
logreg.mc.control(nburn=1000, niter=25000, hyperpars=0, update=0,
output=4)
Arguments
| nburn | number of burn in MCMC iterations that are ignored when computing summaries |
| niter | number of MCMC iterations that are used to compute summary statistics |
| hyperpars | hyperparameters. The code allows up to 10 such parameters, but currently only one is used. In particular, log(P(size=k)/P(size=k+1)) equals hyperpars[1], where P is the prior on model size. Since a maximum model size (specified in logreg is being used, hyperpars[1] can even be smaller than 0. |
| update | every how many iterations there should be an update of the scores. I.e. if update = 1000, a score will get printed every 1000 iterations. So if iter = 100000 iterations, there will be 100 updates on your screen. If update = 0, a one line summary for each fitted model is printed. If update = -1, there is virtually no printed output. |
| output | If abs(output) > 1 bivariate statistics are gathered, if abs(output) > 2 trivariate statistics are also gathered, otherwise only univariate statistics are gathered. If output > 0 all fitted models are saved in a text file “slogiclisting.tmp”, if output < 0 this does not happen. |
Details
Considerations for setting nburn and niter are as for any MCMC problem. In our experience Logic Regression mixes quickly, and a real small nburn(1000, for example) suffices. If there are many trees and large models niter may need to be large.
A more detailed description of the output options can be found in the helpfile of logreg.
Value
A list with arguments nburn, niter, hyperpars, update, and output, that can be used as the value of the argument mc.control of logreg.
Author(s)
Ingo Ruczinski (ingo@jhu.edu) and Charles Kooperberg (clk@fredhutch.org).
References
Ruczinski I, Kooperberg C, LeBlanc ML (2003). Logic Regression, Journal of Computational and Graphical Statistics, 12, 475-511.
Ruczinski I, Kooperberg C, LeBlanc ML (2002). Logic Regression - methods and software. Proceedings of the MSRI workshop on Nonlinear Estimation and Classification (Eds: D. Denison, M. Hansen, C. Holmes, B. Mallick, B. Yu), Springer: New York, 333-344.
Kooperberg C, Ruczinki I (2005). Identifying interacting SNPs using Monte Carlo Logic Regression, Genetic Epidemiology, in press.
See Also
logreg, logreg.tree.control, logreg.anneal.control
Examples
mymccontrol <- logreg.mc.control(nburn = 500, niter = 500000, update = 25000,
hyperpars = log(2), output = -2)