Behavioral and sentiment divers of green agricultural inputs adoption among Chinese farmers

Based on the practical challenges China faces in promoting the green transformation of agriculture and the low adoption rate of green agricultural inputs, this study utilizes the online review data of e-commerce platforms to investigate the factors influencing farmers’ adoption decisions. This study employed Latent Dirichlet Allocation (LDA) to identify topics and KNN to analyze the sentiment trends. This study found that the factors influencing farmers’ decisions can be classified into six themes: product investment return perception (23.24%), product improvement credibility perception (20.24%), crop growth needs perception (16.68%), price and service value perception (14.64%), delivery guarantee perception (14.03%) and product reputation evaluation perception (11.17%). They can be divided into three layers: the basic guarantee layer, the effect verification layer, and the trust building layer. These layers follow a progressive logic, from basic guarantee to effect verification, and finally to building trust. Furthermore, in the analysis of negative emotions under each topic, it is found that the topic with the most negative emotions is product reputation evaluation perception, and the F.P.E.T. anchor framework for negative emotion judgment is proposed. This study theoretically summarizes that farmers’ online decision-making is a progressive cognitive process based on risk aversion and constructs an anchor framework for diagnosing trust collapse. These findings provide a theoretical basis for the design of an intelligent recommendation system for green agricultural inputs and offer practical insights into marketing and trust-building strategies for green agricultural inputs.