Figures
About This Image
In many applications, Bayesian methods are increasingly recognized to yield high-quality results, while providing an accurate measure of uncertainty in the form of posterior distributions. However, they have also obtained a reputation of requiring enormous computational effort and being difficult to use, due to the expertise required in choosing prior distributions. In this article, we focus on Bayesian inference in Hidden Markov Models, which are central to segmentation tasks such as identification of Copy Number Variants (CNV). We achieve dramatic improvements in both speed and accuracy through the use of dynamic wavelet compression and automatically selected priors. Wiedenhoeft et al.
Image Credit: Wiedenhoeft et al.
Citation: (2016) PLoS Computational Biology Issue Image | Vol. 12(5) May 2016. PLoS Comput Biol 12(5): ev12.i05. https://doi.org/10.1371/image.pcbi.v12.i05
Published: May 31, 2016
Copyright: © 2016 Wiedenhoeft et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
In many applications, Bayesian methods are increasingly recognized to yield high-quality results, while providing an accurate measure of uncertainty in the form of posterior distributions. However, they have also obtained a reputation of requiring enormous computational effort and being difficult to use, due to the expertise required in choosing prior distributions. In this article, we focus on Bayesian inference in Hidden Markov Models, which are central to segmentation tasks such as identification of Copy Number Variants (CNV). We achieve dramatic improvements in both speed and accuracy through the use of dynamic wavelet compression and automatically selected priors. Wiedenhoeft et al.
Image Credit: Wiedenhoeft et al.