| 关鹏,王曰芬,傅柱.融合技术主题互补性和合作网络相似性的企业技术合作伙伴预测研究[J].数字图书馆论坛,2026,22(2):31~42 |
| 融合技术主题互补性和合作网络相似性的企业技术合作伙伴预测研究 |
| IResearch on the Prediction of Enterprise Technology Partners Based on the Complementarity of Technology Topics and the Similarity of Cooperation Networks |
| 投稿时间:2025-12-06 |
| DOI:10.3772/j.issn .1673-2286.2026.02.004 |
| 中文关键词: 技术主题互补性;合作网络相似性;技术合作;预测;专利 |
| 英文关键词: Complementarity of Technology Topics; Similarity of Cooperative Networks; Technology Collaboration; Prediction; Patent |
| 基金项目:本研究得到国家社会科学基金项目“面向AI4S的场景化智慧知识服务框架研究”(编号:24CTQ029)、安徽省哲学社会科学基金项目“‘科学-技术-产业’关联视角下新兴技术预测研究”(编号:AHSKY2025D58)、天津市哲学社会科学规划重点项目“面向创新链产业链深度融合的京津冀专利技术转移影响机制及提升路径研究”(编号:TJTQ24-001)资助。 |
| 作者 | 单位 | | 关鹏 | 巢湖学院经济与法学学院 | | 王曰芬 | 天津师范大学大数据科学研究院 | | 傅柱 | 江苏科技大学经济管理学院 |
|
| 摘要点击次数: 10 |
| 全文下载次数: 18 |
| 中文摘要: |
| 针对专利密集型产业技术合作决策中的信息不对称问题,构建企业技术合作伙伴预测模型,辅助企业技术合作决策。基于专利文献提取技术知识特征,本研究利用共同专利权人关系构建合作网络,通过Node2vec算法量化企业合作网络结构相似性,再应用作者-主题模型抽取技术主题向量,并结合JS散度计算技术主题互补性指标,构建“技术主题互补性+合作网络相似性”企业潜在技术合作伙伴预测模型。光刻机领域的实证研究表明,相较于传统单一合作网络预测方法,融入技术主题互补性的预测方法性能更优,且可有效识别战略协同型、突破创新型、资源整合型和机会探索型4种类型潜在合作伙伴。研究方法可为专利密集型行业提供数据驱动的合作伙伴推荐工具,帮助企业快速识别潜在合作伙伴。 |
| 英文摘要: |
| To address the information asymmetry in technology collaboration decision-making within patent-intensive industries, this study constructs a prediction model for enterprise technology partners, aiming to assist enterprises in making technology collaboration decisions. Based on patent literature, technological knowledge features are extracted, and collaboration networks are built using co-patentee relationships. The Node2vec algorithm quantifies enterprise network structural similarity, while the AT model extracts technology topic vectors combined with JS divergence to calculate the complementarity of technology topics, establishing a dual-dimensional potential partner identification model integrating both technology and network perspectives. Empirical
research in the field of lithography machines shows that this model outperforms traditional single cooperative network prediction methods. It integrates the complementarity of technical themes and can effectively identify four types of potential partners: strategic synergy type, breakthrough innovation type, resource integration type, and opportunity exploration type. The research method can provide data-driven partner recommendation tools for patent-intensive industries, helping enterprises quickly identify potential partners. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |