Generic Prediction Rules for Regime Change in Bimodal Processes

摘要

Several bimodal complex processes having great socio-economic impact exist in nature, yet more often than not it is non-trivial to find prediction rules for their regime change. Generic prediction rules for regime transition and duration of subsequent regime are given for several two-regime models. It is demonstrated that prediction rules are a universal property of a large class of bimodal processes. An effort is made to explore the applicability of generic prediction rules to real-world bimodal processes, namely the Indian summer monsoon rainfall and the Indian Ocean Dipole event. Time series representative of these phenomena are analyzed to establish their bimodality and to show that prediction rules hold well for these time series also.