Identification of Technical R&D Cooperation Partners Based on Heterogeneous Graph Embedding and Link Prediction
Pan Hong1, Wang Ming2, Zhao Kai2, Zhai Liang2, Liang Guoqiang2, Zhai Dongsheng2
1.School of Accountancy, Beijing Wuzi University, Beijing 101149 2.College of Economics and Management, Beijing University of Technology, Beijing 100124
摘要技术研发合作伙伴识别是提升创新绩效的关键。针对现有方法中技术研发主体属性刻画有限、合作矩阵稀疏导致识别精度受限的问题，本文提出一种融合异质图嵌入与链路预测的技术研发合作伙伴识别模型（technology development partner identification integrating heterogeneous graph embedding link prediction，HGE-LP-TDPI）。首先，构造技术研发合作伙伴识别领域本体结构，基于大语言模型（large language model，LLM）对专利说明书进行语义抽取，生成包含技术研发主体多维技术关联的异质图。其次，设计表征技术关联的元路径，通过LSTM（long short-term memory）时序编码器与多级注意力机制聚合语义信息。最后，基于链路预测算法识别合作伙伴。电化学储能领域的实证研究结果表明：第一，HGE-LP-TDPI模型在AUC（area under curve）值等关键指标上显著优于基准模型，AUC值达95.62%，验证了多维技术关联融合与异质图嵌入对解决数据稀疏性问题的有效性；第二，元路径权重分析揭示了技术问题驱动是合作形成的核心因素，应用领域与技术功效的影响次之；第三，属性消融实验结果表明，技术维度和知识维度属性对本文模型结果贡献度最高。
1 马宏业, 冯南平, 杨善林. 创新生态系统视角下“卡脖子”技术潜在攻关主体识别研究[J]. 情报学报, 2024, 43(11): 1310-1321. 2 操友根, 任声策, 杜梅. 具有全球竞争力的开放创新生态建设[J]. 科学学研究, 2024, 42(9): 1979-1987. 3 王美香, 陈智, 张耀伟, 等. 产业链与创新链融合闭环构建逻辑——基于万华化学的案例研究[J]. 科研管理, 2025, 46(1): 83-94. 4 王龙伟, 任胜钢, 谢恩. 合作研发对企业创新绩效的影响研究——基于治理机制的调节分析[J]. 科学学研究, 2011, 29(5): 785-792. 5 谢小东, 吴洁, 盛永祥, 等. 基于动态图卷积网络的发明人潜在合作伙伴识别方法研究[J]. 数据分析与知识发现, 2025, 9(3): 83-95. 6 谢小东, 吴洁, 盛永祥, 等. 基于图卷积网络的发明人跨领域合作伙伴识别方法[J]. 情报杂志, 2024, 43(4): 175-183, 167. 7 Luo Y N, Zhao X B, Zhou J T, et al. A network integration approach for drug-target interaction prediction and computational drug repositioning from heterogeneous information[J]. Nature Communications, 2017, 8: Article No.573. 8 何喜军, 吴爽爽, 武玉英, 等. 基于属性异构网络表示学习的专利交易推荐[J]. 情报学报, 2022, 41(11): 1214-1228. 9 Lin M, Wang T, Zhu Y F, et al. A heterogeneous directed graph attention network for inductive text classification using multilevel semantic embeddings[J]. Knowledge-Based Systems, 2024, 295: 111797. 10 谢小东, 吴洁, 盛永祥, 等. 基于异构图卷积网络的专利发明人潜在合作伙伴识别方法研究[J]. 现代情报, 2025, 45(6): 14-25. 11 周志刚, 窦路遥, 李毅. 专利视域下融合协同过滤与链路预测的企业潜在合作关系预测研究[J]. 情报杂志, 2023, 42(12): 144-152. 12 琚春华, 诸惠, 曹倩雯. 基于多元融合的企业技术创新合作预测方法研究[J]. 中国科技论坛, 2024(3): 108-119. 13 Chung J, Ko N, Yoon J. Inventor group identification approach for selecting university-industry collaboration partners[J]. Technological Forecasting and Social Change, 2021, 171: 120988. 14 Zhang J, Baden-Fuller C. The influence of technological knowledge base and organizational structure on technology collaboration[J]. Journal of Management Studies, 2010, 47(4): 679-704. 15 Yan Y, Guan J C. How multiple networks help in creating knowledge: evidence from alternative energy patents[J]. Scientometrics, 2018, 115(1): 51-77. 16 Li E Y, Liao C H, Yen H R. Co-authorship networks and research impact: a social capital perspective[J]. Research Policy, 2013, 42(9): 1515-1530. 17 俞荣建, 胡峰, 陈力田, 等. 知识多样性、知识网络结构与新兴技术创新绩效——基于发明专利数据的NBD模型检验[J]. 商业经济与管理, 2018(10): 38-46. 18 Guan J C, Liu N. Exploitative and exploratory innovations in knowledge network and collaboration network: a patent analysis in the technological field of nano-energy[J]. Research Policy, 2016, 45(1): 97-112. 19 曹兴, 赵倩可, 许羿. 新兴技术合作创新网络形成影响因素研究——基于虚拟现实技术的专利数据[J]. 科学决策, 2024(2): 62-78. 20 Hagedoorn J, Lokshin B, Zobel A K. Partner type diversity in alliance portfolios: multiple dimensions, boundary conditions and firm innovation performance[J]. Journal of Management Studies, 2018, 55(5): 809-836. 21 Yayavaram S, Ahuja G. Decomposability in knowledge structures and its impact on the usefulness of inventions and knowledge-base malleability[J]. Administrative Science Quarterly, 2008, 53(2): 333-362. 22 Da’u A, Salim N. Recommendation system based on deep learning methods: a systematic review and new directions[J]. Artificial Intelligence Review, 2020, 53(4): 2709-2748. 23 刘晓燕, 张淑伟, 单晓红. 面向技术融合的合作伙伴识别研究[J]. 科学学研究, 2023, 41(12): 2164-2174. 24 杨震宁, 杜双, 侯一凡. 目标期望与实现匹配效应如何影响联盟稳定——对中国高技术产业联盟的考察[J]. 管理世界, 2022, 38(12): 122-142. 25 阮平南, 王文丽, 刘晓燕. 基于多维邻近性的技术创新网络演化动力研究——以OLED产业为例[J]. 研究与发展管理, 2018, 30(6): 59-66. 26 汪雪锋, 付芸, 邱鹏君, 等. 基于SAO分析的R&D合作伙伴识别研究[J]. 科研管理, 2015, 36(10): 19-27. 27 Deng W W, Ma J. A knowledge graph approach for recommending patents to companies[J]. Electronic Commerce Research, 2022, 22(4): 1435-1466. 28 施国良, 张丽丽, 张笑笑, 等. 基于专利数据的产业链视角下企业合作目标预测研究——以新能源汽车产业为例[J]. 情报理论与实践, 2024, 47(8): 120-128. 29 赵健宇, 董振杰, 余乐安, 等. 基于知识-合作相依网络的技术融合预测与潜在合作伙伴识别研究[J]. 系统工程理论与实践, 2025, 45(3): 996-1013. 30 马霄, 邓秋淼, 张红玉, 等. 融合异质图表示学习与注意力机制的可解释论文推荐[J]. 情报学报, 2024, 43(7): 802-817. 31 王菲菲, 林中潭, 吴昆, 等. 多阶元路径引导的异质图神经网络新闻推荐模型[J]. 系统工程理论与实践, 2024, 44(5): 1561-1576. 32 傅俊英, 彭喆, 郑佳, 等. 基于专利异构网络的中小企业潜在合作伙伴研究——以石墨烯领域为例[J]. 情报学报, 2019, 38(4): 391-401. 33 Yang X, Deng C, Liu T L, et al. Heterogeneous graph attention network for unsupervised multiple-target domain adaptation[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2022, 44(4): 1992-2003. 34 毕达天, 张雪, 孔婧媛, 等. 基于异质图注意力网络与多特征融合的跨社交媒体用户识别研究[J]. 情报学报, 2024, 43(10): 1213-1226. 35 Liu W, Xin T. EntroMAGNN: an entropy-driven metapath-based graph neural network for maritime emergency event prediction[C]// Proceedings of the 20th International Conference on Intelligent Computing. Singapore: Springer, 2024: 170-186. 36 Xing X, Zhang X M, Cui J F, et al. Multi-behavior attention mechanisms graph neural networks based on session recommendation[C]// Proceedings of the 35th Chinese Control and Decision Conference. Piscataway: IEEE, 2023: 4213-4217. 37 Bai X, Wu X, Stojkovic I, et al. Leveraging large language models for improving keyphrase generation for contextual targeting[C]// Proceedings of the 33rd ACM International Conference on Information and Knowledge Management. New York: ACM Press, 2024: 4349-4357. 38 Wang L Y, Du Z F, Jiao W X, et al. Benchmarking and improving long-text translation with large language models[C]// Findings of the Association for Computational Linguistics ACL 2024. Stroudsburg: Association of Computational Linguistics, 2024: 7175-7187. 39 韦景竹, 祝培培, 马芷晴. 因果关联驱动下技术演化路径及其事件间影响力研究[J]. 图书情报工作, 2025, 69(8): 57-68. 40 翟东升, 郭程, 张杰, 等. 基于专利的企业潜在研发伙伴推荐方法研究[J]. 数据分析与知识发现, 2017, 1(3): 10-20. 41 Brennecke J, Rank O. The firm’s knowledge network and the transfer of advice among corporate inventors—a multilevel network study[J]. Research Policy, 2017, 46(4): 768-783. 42 Fu X Y, Zhang J N, Meng Z Q, et al. MAGNN: metapath aggregated graph neural network for heterogeneous graph embedding[C]// Proceedings of the Web Conference 2020. New York: ACM Press, 2020: 2331-2341. 43 Qian M H, Zhao M C, Yang J L, et al. A novel approach to enterprise technical collaboration: recommending R&D partners through technological similarity and complementarity[J]. Journal of Informetrics, 2024, 18(4): 101571. 44 赵展一, 钟永恒, 李贞贞, 等. 基于技术匹配与分型优化的企业研发合作伙伴推荐方法研究[J]. 情报学报, 2025, 44(1): 48-60. 45 赵展一, 李贞贞, 钟永恒, 等. 融合专利类别与语义信息的企业潜在技术关系测算方法研究[J]. 情报理论与实践, 2023, 46(3): 200-208. 46 Grover A, Leskovec J. node2vec: scalable feature learning for networks[C]// Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. New York: ACM Press, 2016: 855-864. 47 Dong Y X, Chawla N V, Swami A. metapath2vec: scalable representation learning for heterogeneous networks[C]// Proceedings of the 23rd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. New York: ACM Press, 2017: 135-144. 48 Veli?kovi? P, Cucurull G, Casanova A, et al. Graph attention networks[C/OL]// Proceedings of the 6th International Conference on Learning Representations. Vancouver: ICLR, 2018-02-16. https://openreview.net/pdf?id=rJXMpikCZ. 49 Kipf T N, Welling M. Semi-supervised classification with graph convolutional networks[PP/OL]. V4. arXiv (2017-02-22). https://arxiv.org/pdf/1609.02907. 50 Hamilton W L, Ying R, Leskovec J. Inductive representation learning on large graphs[C]// Proceedings of the 31st International Conference on Neural Information Processing Systems. Red Hook: Curran Associates, 2017: 1025-1035. 责任编辑 魏瑞斌）
