Estimating manufacturing value-added via hybrid ANFIS-GA and PSO optimization: A macro-analysis of European OECD economies

OpenAlex İngilizce

In recent years, artificial intelligence-based hybrid methods have gained increasing attention for addressing complex and nonlinear problems in engineering and economic systems. This study proposes an innovative approach to estimate manufacturing value-added at the macro level by integrating innovation, entrepreneurship, and environmental indicators. An Adaptive Neuro-Fuzzy Inference System (ANFIS) model is optimized using Genetic Algorithm (GA) and Particle Swarm Optimization (PSO) to improve estimation accuracy. The proposed hybrid models are tested across different population sizes and benchmarked against the standard ANFIS model. Comparative performance evaluation reveals that the ANFIS-GA model with a population size of 25 outperforms other models, achieving the most consistent and accurate estimation results with R²= 0.9080, MAE= 0.1455, MSE= 0.0325, RMSE= 0.1801, and PBIAS= 0.7960. The findings demonstrate the robustness and applicability of the ANFIS-GA model for manufacturing value-added prediction, offering valuable insights for policy makers and industrial decision-makers in enhancing production performance and sustainable development strategies.