Optimizing Nonlinear Predictions With Neural Network Models For Socio-Economic Development

Abstract

This study develops a neural network framework to optimize nonlinear predictions of socio-economic development driven by advanced propulsion technologies and associated infrastructure. Utilizing a panel dataset, the research employs Pearson correlation, principal component analysis (PCA), linear regression, artificial neural networks (ANN, R²=0.993), and multivariate analysis of variance (MANOVA) with Tukey HSD, standardizing data via z-scores. Findings reveal a strong negative correlation (-0.84) between infrastructure access and socio-economic disparities, with ANN outperforming linear regression in predictive accuracy. The framework highlights the role of propulsion-related infrastructure in reducing economic gaps, supporting sustainable development goals. By leveraging advanced computational methods, this study offers a scalable tool for policy optimization in socio-economic development.