基于复杂网络的产业核心技术识别预测研究——以能源产业为例

摘?要：立足情报研究视角，探索建立科学完整的科技文献信息挖掘及核心技术识别预测方法体系，揭示战略前沿领域的核心技术，明确核心技术的重点研发方向和潜在研发趋势。在界定核心技术概念及类型的基础上，利用文献知识聚类识别热点技术，通过RAKE算法抽取技术主要研究内容，以各项热点技术为节点构建复杂网络，通过节点二次聚类和可视化方法展现技术结构网络，采用结构洞理论分析网络和节点特性，从中遴选出关键技术；利用复杂网络链路预测算法，预测技术结构网络中的缺失边产生连接的可能性，判断现有热点技术未来的交叉融合，以此识别潜在新兴技术。以能源产业为例开展实践，对该领域的科技文献开展深入挖掘，揭示领域核心技术，初步验证了方法的可操作性和有效性。

Research on the identification and prediction of industrial core technologies based on complex networks: Taking the energy industry as an example

Abstract: Based on the perspective of information research, this paper tries to establish a scientific and complete S&T literature information mining and core technology identification method, in order to reveal the core technology in the strategic frontier field, and clarify the key R&D direction and potential R&D trend of core technology. On the basis of the definition, types and concepts of core technology, this paper uses literature knowledge clustering technology to identify heated technologies, extracts the main research content of the technology through RAKE algorithm, constructs a complex network with hot technologies as nodes, presents the technical structure network through nodes secondary clustering and visualization method, analyzes the network and node characteristics by using the structure hole theory, and selects the key technologies. Using the link prediction algorithm of complex network, the possibility of connection generated by missing edges in the technical structure network is predicted, and the future cross fusion of existing heated technologies is judged, so as to identify potential emerging technologies. Taking the energy industry as an example to carry out the practice, the in-depth excavation of S&T Literature in this field is carried out to reveal the core technology of the field, and the operability and effectiveness of the method are preliminarily verified.