Time Series Forecasting in the Digital Age: Leveraging Automated Regression Techniques

Abstract

In the realm of time series data analysis, irregular seasonal patterns pose a challenge to accurate modeling. Ishtiaq (2013) introduced an adapted Lucey approach, based on regression modeling, aimed at mitigating this irregularity. However, this method relied heavily on subjective judgment in defining seasons, rendering it less accessible to less experienced users and potentially leading to erroneous conclusions. In response, we propose an "Automated Adaptive Lucey Approach," a computer-programming-based mechanism that replaces subjective judgment with objective criteria outlined in previous research. This approach transforms the subjective nature of season definition into an objective process, enhancing the reliability and accessibility of the modeling technique. To evaluate the effectiveness of our proposed procedure, we applied it to multiple datasets and observed satisfactory results across three illustrative examples presented herein. The outcomes underscore the appropriateness and robustness of the automated adaptive approach, as evidenced by the final results obtained. By eliminating subjective elements and introducing automated criteria, our method enhances the objectivity and usability of seasonal time series analysis, facilitating more accurate and reliable modeling outcomes.