基于BERTopic的无人配送公众认知主题挖掘与演化分析

Public Perception of Unmanned Delivery: Topic Mining and Evolutionary Analysis Using BERTopic

摘要: 在新质生产力加速重构物流产业格局的背景下，把握公众对无人配送技术的深层认知结构与演化规律，对于推动人工智能与现代物流的深度融合具有重要意义。本研究整合抖音与微博评论数据，采用基于Transformer的BERTopic动态主题模型解构公众认知与演化特征。研究结果显示，模型识别出语音技术、生鲜运输等核心议题，并通过层次聚类将其整合为经济效益与社会伦理、配送场景与服务质量、技术本体感知与创新评价、环境交互与安全适应四大维度。演化路径上，揭示了议题从技术储备与舆论平稳期、示范场景先行与应用探索期、路权开放与商业化探索调整期，向新质生产力爆发期跃迁的范式转移规律。本研究结论可为无人配送的技术范式迭代、场景生态构建及产业政策优化提供数据驱动的实证支撑。

Abstract: Against the backdrop of new quality productive forces accelerating the reshaping of the logistics industry landscape, understanding the deep cognitive structure and evolutionary patterns of public perception regarding unmanned delivery technology is crucial for promoting the deep integration of artificial intelligence and modern logistics. This study integrates comment data from Douyin and Weibo platforms and employs the Transformer-based BERTopic dynamic topic model to deconstruct the characteristics of public perception and its evolution. The results indicate that the model identifies core topics such as voice technology and fresh food transportation. Through hierarchical clustering, these topics are synthesized into four dimensions: Economic Benefits and Social Ethics, Delivery Scenarios and Service Quality, Technological Ontology Perception and Innovation Evaluation, and Environmental Interaction and Safety Adaptation. Regarding the evolutionary path, the study reveals a paradigm shift where topics transition from the “Technical Reserve and Public Opinion Stability Period,” through the “Demonstration Scenario Precedence and Application Exploration Period” and the “Right-of-Way Opening and Commercial Exploration Adjustment Period,” ultimately leaping toward the “New Quality Productive Forces Outbreak Period.” The findings provide data-driven empirical support for the technological paradigm iteration, scenario ecosystem construction, and industrial policy optimization of unmanned delivery.