Assessing quantification for the EMS algorithm (Q1338498)

A framework is developed for assessing the potential of the EMS (estimate, maximize and smooth) algorithm to correctly quantify the emission. At first, the EM and EMS algorithms are defined, the existence theory described, and an interpretation of the EMS equations as a weighted averaging procedure is given. The proposed procedure for assessing quantification is given in Section 3. The reinterpretation of quantification in terms of exclusion regions is explained. Initially, (linear) Perron-Frobenius theory is used to obtain uniform estimates for a quantification measure defined in terms of the reconstruction range. Associated exclusion regions are then identified for the (nonlinear) EMS algorithm. Moreover, these estimates show how the smoothing and model matrix affect the size of the exclusion regions and produce a global compression of the reconstruction range. It is shown that, for primitive smoothing, this global compression occurs locally and therefore limits the extent to which rapid changes can be reconstructed. Finally, a pragmatic interpretation of the earlier theoretical results is given in order to assist with the application of the EMS algorithm to practical situations.