Algorithmic Support for Auto-modes of adaptive short-term Forecasting in predictive Analytics Systems

The problem of improving the algorithmic base of predictive analytics systems by algorithmizing the process of parametric adjustment of the Brown’s predictive model is solved. The relevance of this task is due to the need to implement auto-modes for predictive evaluation of business-critical parameters. An analysis of literary sources is carried out, during which the shortcomings of the search approach to solving the problem of parametric synthesis of predictive models are identified. An algorithm for parametric setting of the Brown’s model based on the analysis of the root portrait of retrospective equations is proposed. The concept of a complex analogue of the real range of admissible values of the smoothing parameter in the Brown’s model is introduced. The proposed algorithm can be used both in the classical and in the extended Brown’s model.