Comparative Analysis of Artificial Intelligence Models in Forecasting Regional Economic Potential (The Case of Surkhandarya Region)

This article examines the problem of forecasting regional economic potential on the basis of statistical data for the districts of Surkhandarya region. Within the framework of the study, the forecasting performance of the MLP, Random Forest, XGBoost and LSTM models is assessed through a comparative benchmark analysis. The empirical analysis draws on a panel database of 184 observations covering the period 2011–2023. The models were evaluated using training and test samples separated chronologically. Based on the results, the model with the highest forecasting accuracy was identified, and scientifically grounded proposals and recommendations aimed at developing the economic potential of the region were formulated.