Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{190865,
        author = {VISHNU N.DABHADE and Bharat R. Naiknaware and Akshay P.Deshpande},
        title = {Improving Sales Forecast Accuracy Using Deep Learning Models: A Comparative Study of LSTM, GRU, and Transformers.},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {12},
        number = {8},
        pages = {5485-5491},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=190865},
        abstract = {Accuracy in sales forecasting is an essential element of effective business planning and strategic decision-making. Existing conventional and deep learning models have difficulty in handling the non-linear sales trends, seasonality, and prolonged sales cycles existing in actual sales data. In light of these challenges, this paper seeks to examine the efficacy of deep learning models in enhancing sales forecast accuracy by comparing the efficacy of Long Short-Term Memory (LSTM), Gated Recurrent Unit (GRU), and Transformer deep learning models in sales forecasting based on sales data supplemented by additional influence variables like promotions, prices, and seasonality. The different sales data supplemented by additional influence variables, along with a comparative study of various deep learning models, enable the creation of a comprehensive and uniform research framework. In addition to these, extensive data pre-processing, normalization, and sequence creation strategies are employed in this study to ensure a level playing field. The efficiency of the discussed deep learning models is determined based on conventional sales forecasting accuracy metrics, which include Mean Absolute Error (MAE), Root Mean Square Error (RMSE), Mean Absolute Percentage Error (MAPE), and Coefficient of determination (R²) of actual and forecast sales. The experimental evaluation reveals that, although LSTM and GRU models.},
        keywords = {Sales forecasting, extreme gradient boosting, Auto-regressive Integrated Moving Average, support vector machines, machine learning, and long- and short-term memory.},
        month = {January},
        }