基于专利知识流网络和图表示学习的技术机会链路预测研究

Research on Technology Opportunity Link Prediction Based on Patent Knowledge Flow Network and Graph Representation Learning

[目的/意义]准确地对技术机会进行预测,能够帮助组织评估和发现具有前景的技术机会,支撑和优化技术研发战略决策与布局,从而提升组织在技术创新领域的竞争优势。[方法/过程]提出一种基于VGAE框架的ss-VGAE技术预测方法。首先,运用生长曲线划分和确定目标领域的生命周期,运用能够兼顾主副IPC的方法在各阶段建立专利知识流网络,并分别提取网络中各个节点的语义特征和结构特征;其次,使用变分图自编码器(VGAE)学习节点的向量表示;最后,利用重构出的邻接矩阵进行技术机会链路预测。[结果/结论]使用石墨烯专利数据对本文方法的有效性进行验证,结果显示基于ss-VGAE的技术机会链路预测模型取得的AUC、AP均优于其他模型,并在当前石墨烯知识流网络上的技术机会进行预测,验证该模型能够有效地挖掘潜在的技术机会,支撑技术研发战略决策和组织技术竞争力的提升。

[Purpose/significance]Accurately predicting technological opportunities can help organizations assess and discover promising technological opportunities,support and optimize strategic decision-making and layout of technological R&D,and thus enhancing the organization’s competitive advantage in the field of technological innovation.[Method/process]The ss-VGAE technology prediction method based on the VGAE framework is proposed in this paper.First,we leverage the growth curve to divide and determine the life cycle of field development,and then use the method that can take into account the primary and secondary IPC to establish a patent knowledge flow network at each stage,with semantic features and structural features extracted for each node of the patent knowledge flow networks.Then,the variational graph autoencoder(VGAE)is used to learn the representation of every node.Finally,the reconstructed adjacency matrix is used to predict the technical opportunity.[Result/conclusion]Graphene patent dataset is used to verify the effectiveness of the method in this paper.The results show that the AUC and AP obtained by the link prediction model based on ss-VGAE are better than other models,and it is more effective than other models on the current graphene knowledge flow network.The prediction of technological opportunities shows that our model can effectively tap potential technological opportunities,support strategic decision-making in technological research and development,meanwhile improve organizational technological competitiveness.