Neural network models for statistical analysis and tax planning in agrarian economy

Modern agricultural economy faces a number of challenges, including climate change, market volatility and the need to improve management efficiency. In the context of digitalisation, the introduction of intelligent tools, in particular neural network models for statistical analysis and tax planning, is of particular importance. This study provided an overview of current approaches to the application of artificial neural networks (ANNs) in the agricultural sector. MLP, CNN, RNN, RNN, LSTM architectures and their hybrid variants used for yield forecasting, tax burden estimation, subsidy planning and financial risk analysis were considered. The aim of this study was to identify the potential of neural network models to improve the efficiency of statistical analysis and tax planning in the agricultural economy, as well as to determine their role in the formation of digital management decision support systems. The paper systematised the main areas of application of neural networks in agrarian economy, gave examples of effective solutions and substantiated the practical significance of ANNs for decision support under uncertainty. Special attention was paid to the integration of ANNs into digital platforms of the agrarian sector and the formation of intelligent systems to support fiscal management. The analysis confirmed the high adaptability and forecasting accuracy of neural networks and emphasised the need to develop digital infrastructure and regulatory framework for their widespread implementation. The results of the study can be used in developing strategies for sustainable agricultural development and improving the economic sustainability of agricultural enterprises. The work is of interest to researchers, developers of digital solutions and specialists in the field of agricultural policy