基于BERTopic的电子商务技术主题挖掘与时间演化分析

Research on E-Commerce Technology Topic Mining and Temporal Evolution Analysis Based on BERTopic

摘要: 本研究以2015~2024年间中国国家知识产权局收录的5750项电子商务专利为分析对象，运用BERTopic深度主题模型对该领域技术主题进行挖掘与演化分析。研究识别出10个技术主题，根据其时间演化特征归纳为四大特征集群：“生命周期型”硬件自动化、“达峰缓降型”核心平台、“稳态关注型”商品信息化及“低频成熟型”综合支撑。时序分析显示，电子商务技术呈现出从基础建设到智能化发展的阶段性特征。基于上述发现，本文从企业差异化技术投资与产业生态培育两个层面提出了优化路径建议。研究为电子商务技术演进提供了量化分析框架，并对未来沉浸式交互、全局智能化及规范化治理等发展方向进行了展望。

Abstract: This study analyzes 5750 e-commerce patents from the China National Intellectual Property Administration from 2015 to 2024, employing the BERTopic deep topic modeling approach to mine and analyze technological themes and their evolution. Ten distinct technological topics were identified and categorized into four characteristic clusters based on their temporal evolution patterns: “life-cycle-type” hardware automation, “peak-and-decline-type” core platforms, “steady-state-type” product informatization, and “low-frequency-maturity-type” comprehensive support technologies. The temporal analysis reveals that e-commerce technology development exhibits phased characteristics, progressing from foundational construction to intelligent advancement. Based on these findings, optimization pathways are proposed from the perspectives of differentiated corporate technology investment and industrial ecosystem cultivation. The study provides a quantitative analytical framework for understanding e-commerce technology evolution and offers insights into future development directions such as immersive interaction, global intelligentization, and standardized governance.