Estimating the parameters for an exponential process via a Hough transform

A technique for estimating the decay rate and predicting the steady-state level of an exponential process is developed. For a time-varying process that is not purely exponential, the technique produces moving estimates and predictions. The technique gives a differential model of an exponential with a Hough transform. The differential model generalizes the definitions of decay rate and steady-state level to encompass time variation, constraining one in terms of the other. The Hough transform then disentangles the two, using a continuity assumption to solve an otherwise ill-posed problem. The technique is computationally simpler than maximum likelihood and compares favorably to it when applied to real data