EUVIMEDEUVIMED entdecken
Crossref

A Method for Adaptive Classification of Goods Based on Forecasted Characteristics to Support Decision-Making in Inventory Management

Kovtun V

Artificial Intelligence · 30.06.2026 · 31 (AI.2026.31(2)) · 97-110

DOI: 10.15407/jai2026.02.097 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Artificial Intelligence, erschienen am 30.06.2026. Verfasst von Kovtun V. Die Arbeit nennt 30 Literaturangaben.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

The article addresses the problem of improving decision support in inventory management under conditions of demand uncertainty, expanding product assortments, and increasing market dynamics. It is shown that traditional product classification approaches, particularly ABC-XYZ analysis, are primarily based on historical data and do not account for potential changes in product characteristics in future periods. This limits the ability to respond promptly to demand fluctuations and reduces the effectiveness of inventory management. The purpose of this study is to develop a method for adaptive product classification based on forecasted product characteristics to support inventory management decision-making. The proposed approach combines a multi-factor representation of product items, forecasting of their characteristics using machine learning methods, and a mechanism for adaptive review of product assignment to ABC-XYZ classes. The study develops a system of product characteristics that includes indicators of economic value, demand parameters, inventory availability indicators, and logistics-related characteristics. A distinctive feature of the proposed approach is the consideration of inventory availability factors, which helps avoid misinterpretation of declining sales caused by stockouts rather than by a decrease in actual demand. Unlike traditional approaches, the proposed method uses forecasted product characteristics to determine the future state of a product item and to form its forecast class. To evaluate the necessity of reviewing a product’s class assignment, an integrated indicator of product state change is proposed. This indicator incorporates forecasted changes in all product characteristics and reduces the influence of random fluctuations in individual indicators. Class review is performed only when differences exist between the current and forecast classes and when the integrated indicator exceeds a predefined threshold. The proposed approach employs configurable class-review thresholds, enabling adaptation to different industries, assortment policies, and operating conditions of enterprises. The scientific novelty of the study lies in the development of a method for adaptive product classification that provides automatic review of product assignment to ABC-XYZ classes based on forecasted product characteristics and an integrated assessment of changes in product state. The practical significance of the proposed approach lies in improving the validity of managerial decisions related to inventory control, monitoring, and replenishment, as well as creating the prerequisites for the transition from reactive to proactive inventory management under demand uncertainty.

Zitieren

APA

V, K. (2026). A Method for Adaptive Classification of Goods Based on Forecasted Characteristics to Support Decision-Making in Inventory Management. Artificial Intelligence, 31(AI.2026.31(2)), 97-110. https://doi.org/10.15407/jai2026.02.097

Vancouver

V K. A Method for Adaptive Classification of Goods Based on Forecasted Characteristics to Support Decision-Making in Inventory Management. Artificial Intelligence. 2026;31(AI.2026.31(2)):97-110. doi:10.15407/jai2026.02.097

BibTeX
@article{V2026,
  author = {V, Kovtun},
  title = {A Method for Adaptive Classification of Goods Based on Forecasted Characteristics to Support Decision-Making in Inventory Management},
  journal = {Artificial Intelligence},
  year = {2026},
  volume = {31},
  number = {AI.2026.31(2)},
  pages = {97--110},
  doi = {10.15407/jai2026.02.097}
}
Kostenlos starten →