MC(4): a tempering algorithm for large-sample network inference
Barker, D., Hill, Steven (Steven M.) and Mukherjee, Sach (2010) MC(4): a tempering algorithm for large-sample network inference. In: 5th International Conference on Pattern Recognition in Bioinformatics, Nijmegen, The Netherlands, 22-24 Sep 2010. Published in: Pattern Recognition in Bioinformatics, Vol. 6282 pp. 431-442.Full text not available from this repository.
Bayesian networks and their variants are widely used for modelling gene regulatory and protein signalling networks. In many settings, it is the underlying network structure itself that is the object of inference. Within a Bayesian framework inferences regarding network structure are made via a posterior probability distribution over graphs. However, in practical problems, the space of graphs is usually too large to permit exact inference, motivating the use of approximate approaches. An MCMC-based algorithm known as MC(3) is widely used for network inference in this setting. We argue that recent trends towards larger sample size datasets, while otherwise advantageous, call, for reasons related to concentration of posterior mass, render inference by MC(3) harder. We therefore exploit an approach known as parallel tempering to put forward an algorithm for network inference which we call MC(4). We show empirical results on both synthetic and proteomic data which highlight the ability of MC(4) to converge faster and thereby yield demonstrably accurate results, even in challenging settings where MC(3) fails.
|Item Type:||Conference Item (Paper)|
|Divisions:||Faculty of Science > Centre for Complexity Science|
|Journal or Publication Title:||Pattern Recognition in Bioinformatics|
|Page Range:||pp. 431-442|
|Conference Paper Type:||Paper|
|Title of Event:||5th International Conference on Pattern Recognition in Bioinformatics|
|Type of Event:||Conference|
|Location of Event:||Nijmegen, The Netherlands|
|Date(s) of Event:||22-24 Sep 2010|
Actions (login required)