Complex exponential smoothing

Exponential smoothing has been one of the most popular forecasting methods used to support various decisions in organizations, activities such as inventory management, scheduling, revenue and other areas. Although its relative simplicity transparency have made it very attractive for research practice, identifying underlying trend remains challenging with significant impact on resulting accuracy. This resulted development modifications models, introducing a model selection problem. With aim addressing this problem, we propose complex exponential (CES), based theory functions variables. The basic CES approach involves only two parameters does not require procedure. Despite these simplifications, proves be competitive with, or even superior existing methods. We show that several advantages over conventional models: can forecast both stationary non-stationary processes, capture level cases, defined classification. is evaluated competition datasets, demonstrating better performance than established benchmarks. conclude desirable features time series modeling opens new promising avenues research.