Maximum entropy analysis of oversampled data problems
- 574 Downloads
An algorithm for the solution of the Maximum Entropy problem is presented, for use when the data are considerably oversampled, so that the amount of independent information they contain is very much less than the actual number of data points. The application of general purpose entropy maximisation methods is then comparatively inefficient. In this algorithm the independent variables are in the singular space of the transform between map (or image or spectrum) and data. These variables are much fewer in number than either the data or the reconstructed map, resulting in a fast and accurate algorithm. The speed of this algorithm makes feasible the incorporation of recent ideas in maximum entropy theory (Skilling 1989 a; Gull 1989). This algorithm is particularly appropriate for the exponential decay problem, solution scattering, fibre diffraction, and similar applications.
Key wordsMaximum entropy Inverse problem Dynamic light scattering
Unable to display preview. Download preview PDF.
- Bryan RK (1980) Maximum entropy image processing. PhD Thesis, University of CambridgeGoogle Scholar
- Golub GH, van Loan CF (1983) Matrix computations. Johns Hopkins, BaltimoreGoogle Scholar
- Gull SF (1989) Developments in maximum entropy data analysis. In: Skilling J (ed) Maximum entropy and Bayesian methods. Kluwer, Dordrecht, pp 53–71Google Scholar
- Gull SF, Skilling J (1984) Maximum entropy method in image processing. IEE Proc 131(F):646–659Google Scholar
- Jaynes ET (1968) Prior probabilities. IEEE Trans SCC-4:227–241Google Scholar
- Jaynes ET (1987) Bayesian spectrum and chirp analysis. In: Smith C Ray, Erickson GJ (ed) Maximum entropy and Bayesian spectral analysis and estimation poblems. Kluwer, Dordrecht, pp 1–37Google Scholar
- Pollard KOB (1988) Thesis, University of CambridgeGoogle Scholar
- Sibisi S (1989) Regularization and inverse problems. In: Skilling J (ed) Maximum entropy and Bayesian methods. Kluwer, Dordrecht, pp 389–396Google Scholar
- Skilling J (1989a) Classic maximum entropy. In: Skilling J (ed) Maximum entropy and Bayesian methods. Kluwer, Dordrecht, pp 45–52Google Scholar
- Skilling J (1989b) The eigenvalues of mega-dimensional matrices. In: Skilling J (ed) Maximum entropy and Bayesian methods. Kluwer, Dordrecht, pp 455–466Google Scholar
- Turchin VF, Nozik VZ (1969) Statistical regularization of the solution of incorrectly posed problems. Atmos Oceanic Phys 5:14–18Google Scholar