Goal-Oriented Time-Series Forecasting: Foundation Framework Design

Abstract: Traditional time-series forecasting often focuses only on minimizing prediction errors, ignoring the specific requirements of real-world applications that employ them. This paper presents a new training methodology, which allows a forecasting model to dynamically adjust its focus based on the importance of forecast ranges specified by the end application. Unlike previous methods that fix these ranges beforehand, our training approach breaks down predictions over the entire signal range into smaller segments, which are then dynamically weighted and combined to produce accurate forecasts within a region of interest. We tested our method on standard datasets, including a new wireless communication dataset, and found that not only it improves prediction accuracy but also enhances the performance of end application employing the forecasting model. This research provides a basis for creating forecasting systems that better connect prediction and decision-making in various practical applications.