Wavelet decomposition is a well-known technique to compress image data. Here, we have used wavelet decomposition to compress digital elevation models, DEM. The objectives for compressing DEMs are obtaining manageable and small data sets, and reducing data access time. In the paper, different aspects of wavelets as a base for data compression are described. The (de-)compression scheme consists o...

Bayesian neural networks (BNNs) with latent variables are probabilistic models which can automatically identify complex stochastic patterns in the data. We study in these models a decomposition of predictive uncertainty into its epistemic and aleatoric components. We show how such a decomposition arises naturally in a Bayesian active learning scenario and develop a new objective for reliable re...

Coupling methods for continuum models with molecular models are developed. Two methods are studied here: an overlapping domain decomposition method, which has overlapping domain, and an edge-to-edge decomposition method, which has an interface between two models. These two methods enforce the compatibility on the overlapping domain or interface nodes/atoms by the Lagrange multiplier method or t...

Multivariate Decomposition for Hazard Rate Models We develop a regression decomposition technique for hazard rate models, where the difference in observed rates is decomposed into components attributable to group differences in characteristics and group differences in effects. The baseline hazard is specified using a piecewise constant exponential model, which leads to convenient estimation bas...

Proper Orthogonal Decomposition (POD), alternatively known as Principal Component Analysis or the Karhunen-Loève decomposition, is a model-reduction technique which generates the optimal linear subspace of dimension D for a given set of higher-dimensional data. That is, if the data are contained within an attractor, the POD process can produce the affine linear space that best approximates the ...

Hydrological time series forecasting remains a difficult task due to its complicated nonlinear, non-stationary and multi-scale characteristics. To solve this difficulty and improve the prediction accuracy, a novel four-stage hybrid model is proposed for hydrological time series forecasting based on the principle of 'denoising, decomposition and ensemble'. The proposed model has four stages, i.e...

Since the decomposition rate of soil organic carbon (SOC) varies as a function of environmental conditions, global climate change is expected to alter SOC decomposition dynamics, and the resulting changes in the amount of CO2 emitted from soils will feedback onto the rate at which climate change occurs. While this soil feedback is expected to be significant because the amount of SOC is substant...

This paper addresses the problem of reliably estimating the coeÆcients of the parameterized image variety (PIV) [3] associated with the set of weak perspective images of a rigid scene, with applications in image-based rendering. Exploiting the fact that the constraints de ning the PIV are linear in its coeÆcients and bilinear in the image data, the estimation procedure is cast in the errors-in-...

This paper presents practical methods for the sequential generation or simulation of a Gaussian two-dimensional random field. The specific realizations typically correspond to geospatial errors or perturbations over a horizontal plane or grid. The errors are either scalar, such as vertical errors, or multivariate, such as x, y, and z errors. These realizations enable simulation-based performanc...