The use of a single continuous regularization parameter for the generalized eigensystem method

We have previously (1994, 1997, 1999, 2000) proposed the generalized eigensystem (GES) method as an alternative to Tikhonov regularization for various types of inverse problems in electrocardiography. In those studies, we truncated the number of modes used in GES to stabilize the inversion. However, often there is a significant decrease in performance when the number of modes to use is estimated using only the available data. We have modified our formulation of GES to use the same expansion vectors as before, but now the stabilization is performed by the inclusion of a single continuous regularization parameter. In this study, we compare the performance of GES using the optimal modal truncation and two different methods for choosing the single continuous regularization parameter