The technique of linear prediction filters applied to studies of solar wind-magnetosphere coupling

Linear prediction filtering is a powerful empirical technique suitable for the study of stimulus-response behavior. The technique enables one to determine the most general linear relationship between multiple time-varying quantities, assuming that the physical systems relating the quantities are linear and time invariant. Several researchers have applied linear prediction analysis to investigate solar wind-magnetosphere interactions. This short review describes the method of linear prediction analysis, its application to solar wind-magnetosphere coupling studies both in terms of physical processes, and the results of investigations which have used this technique.