A Non‐Gaussian Spatial Process Model for Opacity of Flocculated Paper

From MaRDI portal





The authors discuss the class of non-Gaussian stochastic models for continuous spatial variation appropriate to describe paper opacity data, which are determined by on-line measurements with a sensor giving a one-dimensional longitudinal trace along the manufacturing direction. Such models are essential for developing the on-line quality control in paper-making industry. The class of models under consideration is given by a real-valued stochastic process \(Y(x), x\in R^2,\) which admits the decomposition \(Y(x)=Y_1(x)+Y_2(x)\), where \(Y_1(x)=\beta \sum_j f(x-x_j)\), \(f(\cdot)\) is a bivariate probability density, \(x_j\) are points of a stochastic point process in \(R^2\), and \(Y_2(x)\) is a white noise. Here \(Y_1, Y_2\) are independent random processes, \(Y_1\) models the flocculation process and \(Y_2\) is intended to describe micro-scale variation in the opacity of the paper when no flocculation occurs. The authors investigate the spectral properties of the model and develop methods of parameter estimation based on maximum likelihood in the frequency domain. The proposed methods are illustrated using simulated and real data.











This page was built for publication: A Non‐Gaussian Spatial Process Model for Opacity of Flocculated Paper

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4828204)