| 边竞 白晨 曹悦 张玄玄 张越 李妙钰.融合生成式主题建模与多源异构数据的技术演化预见研究[J].中国科技资源导刊,2026,(4):55~67 |
| 融合生成式主题建模与多源异构数据的技术演化预见研究 |
| Research on Technical Evolution and Foresight by Integrating Generative Topic Modeling with Multi-source Heterogeneous Data |
| 投稿时间:2026-11-10 |
| DOI: |
| 中文关键词: 技术演化;技术预见;TopicGPT;多源异构数据;智能网联汽车 |
| 英文关键词: technology evolution, technology forecasting, TopicGPT, multi-source heterogeneous data, intelligent connected vehicles |
| 基金项目:中国科学技术信息研究所创新研究基金青年项目“基于生成式主题建模与多源异构数据的技术演化预见研究——以智能网联汽车产业为例”(QN2025-22)。 |
| 作者 | 单位 | | 边竞 白晨 曹悦 张玄玄 张越 李妙钰 | (中国科学技术信息研究所,北京 100038) |
|
| 摘要点击次数: 35 |
| 全文下载次数: 22 |
| 中文摘要: |
| 系统梳理技术演化脉络并预见其未来发展趋势,对于政府规划产业布局、企业锚定创新方向至关重要。本文聚焦智能网联汽车领域,以2015—2024年论文、专利、政策、报告4类文本作为数据源,构建多源异构数据库;采用生成式主题建模工具TopicGPT,搭建融合多源数据特征的技术演化预见框架;识别智能网联汽车领域内技术主题的演化路径与阶段特征,并对关键主题的发展趋势进行外推预测。研究表明:相较传统的方法,本文提出的框架能够更加全面细致地揭示智能网联汽车领域的技术演化机理;趋势预测结果对指明未来潜在技术突破方向具有重要意义。 |
| 英文摘要: |
| Systematically tracing the evolution of technies and anticipating their future trends is crucial for government planning of industrial layouts and for enterprises to anchor their innovation directions. Focusing on the intelligent connected vehicle domain, this study constructs a multi-source heterogeneous database using four types of texts—papers, patents, policies, and reports—from 2015 to 2024. Employing the generative topic modeling tool TopicGPT, we establish a technology evolution forecasting framework that integrates multisource data features. This framework identifies the evolutionary paths and stage characteristics of technical themes within the intelligent connected vehicle domain and extrapolates future trends for key themes. The study demonstrates that compared to traditional approaches, the proposed framework reveals the technical evolution mechanisms of the intelligent connected vehicle domain in a more comprehensive and detailed manner. The trend prediction results hold significant implications for identifying potential future technical breakthrough directions. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |