Comparative analysis of types of neural networks for solving problems of modeling socioeconomic systems (forecasting of production using neural networks, for example, on an LSTM-type network)

This study discusses modern production forecasting methods using neural networks, in particular, the long short-term memory (LSTM) architecture. Given the complexity and dynamic variability of production processes, traditional approaches are often insufficient for accurate planning. Neural networks, especially LSTM, are becoming increasingly popular due to their ability to effectively process time series and take into account long-term dependencies in data. This study presents examples of successful applications of LSTM for demand forecasting, inventory optimization, and higher efficiency of production processes. It also evaluates the advantages of using neural networks compared to classical methods of data collection and analysis, such as regression and autoregressive models. The authors focus on tuning hyperparameters, as well as on the need for high-quality data preprocessing to achieve maximum accuracy. In addition, this study includes an analysis of practical cases illustrating a significant improvement in forecast accuracy and cost reduction after the implementation of neural network approaches. The authors hope that the results of this study will be useful to both researchers and practitioners in the fields of production management and management decision-making.