We describe how constraint propagation techniques can be used to reliably reconstruct model parameters from noisy data. The main algorithm combines a branch and bound procedure with a data inflation step; it is robust and insensitive to noise. The set–valued results are transformed into point clouds, after which statistical properties can be retrieved. We apply the presented method to a mixed-effects model.
|Title of host publication||Proceedings of the 2010 International Symposium on Nonlinear Theory and its Applications (NOLTA2010)|
|Place of Publication||Japan|
|Publisher||Institute of Electronics, Information and Communications Engineers (IEICE)|
|Number of pages||4|
|Publication status||Published - 2010|