Computational model for agricultural vocation in Spain using artificial neural networks, neuro-fuzzy systems, and support vector machines

Agricultural employment plays a critical role in food security and rural sustainability; however, it has experienced a sustained decline, largely driven by labor reallocation toward the service sector in Spain. Despite its relevance for public policy design, existing research has predominantly focused on agricultural production, with a notable lack of studies addressing agricultural employment dynamics using artificial intelligence techniques. This study addresses this gap by proposing a modular computational framework to model agricultural employment in Spain using artificial neural networks (ANN), neuro-fuzzy inference systems (ANFIS), and support vector machines (SVM). The framework integrates demographic evolution, sectoral participation, and labor disaggregation. In addition, model performance is evaluated using standard quantitative error metrics (MSE, MAE, and MAPE) and statistical comparison procedures. The results show that ANN and ANFIS achieve consistently lower prediction errors than SVM across most modules, with statistically significant differences in several outputs (p < 0.05), while ANN provides the highest predictive accuracy in aggregated demographic and participation modules. ANFIS exhibits competitive accuracy while offering intrinsic interpretability through rule-based representations. By balancing quantitative performance and interpretability, this research extends the application of artificial intelligence to agricultural employment modeling and provides actionable, data-driven insights to support the formulation of agricultural and rural development policies, particularly within the European context.