Policy attention dynamics in China’s agricultural and rural modernization: a machine learning approach

As a strategic force in shaping agricultural and rural development, government attention constitutes a pivotal element in the modernization process. This study conducts a comprehensive machine learning-based textual analysis of 3,718 Chinese government policy documents on agricultural and rural modernization issued between 2017 and 2023, to explore the patterns of government attention reflected therein. Specifically, the Latent Dirichlet Allocation (LDA) model is applied to identify major attention topics. In contrast, dynamic topic modeling (DTM), word embedding technique, and the Term Frequency-Inverse Document Frequency (TF-IDF) algorithm are used to analyze temporal and spatial dimensional differences in attention allocation. The results indicate that governmental attention at all administrative levels is closely aligned with the Five-Sphere Integrated Plan, demonstrating a consistent focus and scope orientation. However, as the administrative hierarchy ascends, the diversity of attention topics decreases, whereas the distributional balance in allocation intensity across sub-issues improves. Moreover, local governments with diverse resource endowments exhibit distinctive attentional preferences and a notable neighborhood effect. This study employs machine learning techniques to assess government attention to agricultural and rural modernization, enriching its research methodologies by integrating the historical evolution and horizontal characteristics of attention allocation.