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  

Recommendation Method for Enterprise R&D Partners Using Technology Matching and Typological Optimization

作　　者：赵展一 钟永恒 李贞贞[3,4] 刘佳[2,3,4] 席崇俊 Zhao Zhanyi;Zhong Yongheng;Li Zhenzhen;Liu Jia;Xi Chongjun(National Science Library,Chinese Academy of Sciences,Beijing 100190;Department of Information Resources Management,School of Economics and Management,University of Chinese Academy of Sciences,Beijing 100190;National Science Library(Wuhan),Chinese Academy of Sciences,Wuhan 430071;Hubei Key Laboratory of Big Data in Science and Technology,Wuhan 430071)

机构地区：[1]中国科学院文献情报中心,北京100190 [2]中国科学院大学经济与管理学院信息资源管理系,北京100190 [3]中国科学院武汉文献情报中心,武汉430071 [4]科技大数据湖北省重点实验室,武汉430071

出　　处：《情报学报》2025年第1期48-60,共13页Journal of the China Society for Scientific and Technical Information

基　　金：国家社会科学基金青年项目“市场导向下的存量专利优选与智能供需匹配模型研究”(24CTQ058)。

摘　　要：从众多创新主体中为企业精准推荐与其需求相匹配的合作伙伴,有利于降低创新风险、攻克关键核心技术。为提高研发合作伙伴推荐结果的精准性和可解释性,本文提出一种基于技术匹配与分型优化的企业研发合作伙伴推荐方法。该方法以技术匹配度(包括技术相似度和互补度)为依据,通过机器学习算法识别大规模潜在合作组对,利用波士顿矩阵从技术相似度和互补度两个维度对伙伴类型进行细分,并结合创新实力、合作偏好、邻近程度和品牌效应等指标,对分型结果进行评价,最终筛选出最佳伙伴。以燃料电池领域为例进行实证应用,结果表明,类别-语义维度的识别算法F1值为93%,比基于类别维度、语义维度的识别算法F1值分别高出2个和4个百分点,可以准确反映创新主体间的技术匹配程度;支持将合作伙伴细分为优先合作型、重点关注型、变革补充型和多元扩张型4类;优化后的结果可为企业提供并有效区分多样化的选择,提升合作成功概率。Accurately recommending compatible partners among diverse innovation entities is crucial for reducing innovation risks and overcoming key core technological challenges.This study proposes a method for recommending enterprise R&D partners based on technology matching and typological optimization to enhance precision and interpretability.The method integrates technological matching,including technological similarity and complementarity to identify large-scale potential cooperation pairs using machine learning algorithms.The Boston Matrix is employed to categorize these recommendations across two dimensions:technological similarity and complementarity.Additionally,indicators such as innovation strength,cooperation preferences,proximity,and brand effect are combined to evaluate the typological results and optimize the entire process.Using the field of fuel cells for an empirical application,the results depict that the models identification algorithm achieves an F1 value of 93%,outperforming category and semantic dimensions-based algorithms by 2 and 4 percentage points,respectively.This model accurately reflects technological matching between innovation entities and supports the subdivision of partners into four categories:priority cooperation,key focus,transformative complement,and diversified expansion.The evaluation and optimization results can effectively distinguish and provide various options for enterprises,increasing the likelihood of successful cooperation.

关 键 词：研发合作 技术匹配 技术相似 技术互补 分型优化 

分 类 号：F273.1[经济管理—企业管理]