What Are the Preferences of Chinese Farmers for Drones (UAVs): Machine Learning in Technology Adoption Behavior

With the continuous advancement of sustainable agriculture, drone technology has become a focus of attention. Current research primarily relies on classical models for questionnaire surveys and analyses within specific regions, rather than implementing macro-level investigations that incorporate innovative algorithms. This study designed a survey questionnaire to investigate Chinese farmers’ preferences for agricultural drones and their technology adoption mechanisms under sustainable agriculture context. The Ant Colony Optimization-Decision Tree (ACO-DT) model and SHAP (Shapley Additive exPlanations) value analysis are applied to analyze the contribution of different indicators to technology adoption. The ACO-DT model outperformed traditional machine learning models with approximate accuracy 0.85, recall 0.98, and F1 Score 0.90, effectively identifying potential drone users compared to other traditional machine learning models. The SHAP analysis showed “Time Required for Promotion” (average SHAP value exceeds 1.25) and “Understanding of UAV Agriculture” (average SHAP value is about 1.0) were core influencing factors. Specifically, high-cognition farmers preferred shorter promotion cycles, while low-cognition group favored longer cycles to reduce decision-making uncertainty. Practically, the study enriches agricultural technology adoption research methodologically and offers references for advancing smart agriculture and optimizing rural production factors.