Transformers in Data-Driven Decision-Making: Applications for Forecasting Sales, Analyzing Demand, and Optimizing Pricing Strategies

Аннотация

This article examines the methodological and Economics Innovations, 63–69. https://doi.org/10.37547/tajmei/Volume07Issue03-09 aspects of applying Transformer architectures for sales forecasting, demand analysis, and price optimization. COPYRIGHT The focus is on the development, adaptation, and © 2025 Original content from this work may be used under the terms of the creative commons attributes 4.0 License. integration of models in the context of processing large volumes of data and operating complex market mechanisms. The paper explores approaches to combining time series, identifying factor relationships, and improving the accuracy of analytical conclusions. The methodology includes adapting basic Transformer architectures, such as Transformer with Multihead Attention Mechanism, to the specific characteristics of the data. The preparatory steps cover information aggregation, creation of temporal features, identification of categorical variables, and handling missing data. Historical datasets supplemented with external information sources are used for training. The sources include scientific articles by international authors published in open access, as well as materials available on the internet, allowing for a broad examination of the topic. The results demonstrate the effectiveness of these architectures in forecasting tasks, identifying temporal dependencies, and improving business process quality. Examples of model implementation illustrate their successful use in commercial information systems. The conclusions emphasize the role of these approaches in decision-making automation and strategic planning. The materials of the article are intended for professionals working in machine learning, data The American Journal of Management and Economics Innovations 59 https://www.theamericanjournals.com/index.php/tajmei The American Journal of Management and Economics Innovations analytics, and process management improvement. a hybrid Transformer-BiGRU algorithm was proposed to account for temporal characteristics when processing sales data. The study by Xiang Y. et al. [7] presents a Keywords: sales forecasting, demand modeling, price model that integrates temporal and frequency optimization, time series, machine learning, deep methods to achieve accuracy. Mu S. et al. [3] learning, Transformer architecture. emphasized the flexibility of transformer computations used for sales analysis. Introduction: The current realities of the global market Demand analysis is explored by combining various require the development of accurate demand methods. In the paper by Amellal I. et al. [1], an forecasting algorithms and the implementation of algorithm combining BERT, GRU, and probabilistic adaptive tools for price optimization. Managing these approaches for time series analysis is developed. processes is fundamental to ensuring the Taparia V. et al. [11] describe the integration of competitiveness of companies. Machine learning regression models with machine learning algorithms, methods, including deep learning networks, offer facilitating the processing of retail sales data. The work opportunities for analyzing relationships in large by Smirnov P. S. and Sudakov V. A. [12] presents an datasets that reflect both short-term fluctuations and adaptive method used for predicting demand for new long-term patterns. products. Transformer models, initially designed for processing Price modeling includes the use of flexible algorithms. textual information, have proven to be versatile when A probabilistic Transformer for forecasting electricity working with temporal data and studying interactions prices is described in the work of Celeita Rodriguez D. between various variables. Their adaptation to time F. [5]. Decision Transformer for real-time data analysis series analysis tasks enables the formation of consumer is presented in the paper by Zhang Z. and Wu M. [6]. behavior forecasts, evaluation of marketing strategies' Zhong B. [8] proposed a solution combining LSTM, ANN, impact, and the development of approaches for and Transformer to integrate demand and price demand forecasting and price management. analysis. The use of machine learning technologies facilitates The integration of forecasting with optimization is decision-making automation and improves forecasting discussed in Zhang J. and Zhao J. [10], where sequential accuracy in areas where the information is solutions for managing production, warehouse, and characterized by complex structure, seasonal changes, sales processes are considered. The approach and multi-level dependencies. For the successful highlights the importance of combining analytical implementation of such solutions, it is important to models with management tasks. Li Q. and Yu M. [2] consider data preparation aspects, model parameter developed a sales forecasting model based on a optimization, and the analysis of the results obtained. modified Transformer architecture. The paper focuses This article examine