近年来，创新驱动战略已经成为了我国经济发展的新引擎，各地政府都陆续出台了一系列政策措施来推动科技创新的发展，科技创新政策（以下简称科创政策）发挥了十分重要的引领作用。对科创政策开展计量分析能够让政策的制定者了解当前政策的现状并对政策进行调整。同时也能够让科研单位、企业、个人等科创政策相关群体了解到政策的侧重点和发展态势，以便更好的开展科创产业相关活动，融入到国家发展战略中来。但是科创政策中涉及到的概念要素较多，概念要素之间的语义关联复杂，且传统政策计量的方式比较单一，难以全面深入的对科创政策中各个概念要素及要素之间关系特征展开计量。因此需要一种更加高效的科创政策知识组织方式，以便能够通过政策计量更好的挖掘科创政策概念要素之间的关联以及其蕴含的信息。基于上述需求分析，本文引入了知识图谱技术，致力于构建一个包含了科创政策内外部主要特征和内涵的科创政策图谱，实现科创政策文本资源的语义化组织，并探索科创政策图谱在政策计量领域的应用模式，以便更好的对科创政策进行理解以及更加全面且高效的进行政策计量分析。为实现这个目标，本文将科创政策图谱分为概念层、逻辑层和数据层，并依次进行搭建。首先，本文在吸纳了三螺旋理论、政策工具理论和创新价值链理论等科技创新相关理论后，结合我国的现实情况，对科创政策要素结构进行分析梳理，构建了包含科创政策外部特征、政策工具、科创主体以及创新价值链阶段等核心内容要素的科创政策分析模型，并将此模型作为科创政策图谱的概念层，为科创政策图谱的构建提供理论和概念基础；然后，在该政策分析模型的基础上构建科创政策本体概念模型作为科创政策图谱的知识表示基础，并以粤港澳大湾区深圳和香港两大核心城市的科创政策作为数据样本，根据本体模型的概念类型结构进行科创政策知识的抽取，完成本体实例的创建，实现科创政策图谱逻辑层的构建。接下来，按照本体和图数据库元素之间的映射关系，将抽取到的知识存入到Neo4j图数据库中完成数据层的搭建，并实现了科创政策图谱的构建。随后，本文提出了科创政策图谱在政策计量领域中基于语义关系推理的政策计量关系探索和基于语义关系的计量信息检索两个主要应用模式，梳理了科创政策要素之间的政策计量关系和利用科创政策图谱来检索计量所需信息的方法及流程。最后，本文以深圳和香港两地作为分析对象，选择了5个通过语义关系推理而来的政策计量关系作为计量的角度，并在基于语义关系的信息检索的基础上完成了多维度政策计量分析，验证了科创政策图谱在政策计量领域应用的实用性和可行性。

In recent years, innovation-driven strategy has become a new engine of China's economic development, and local governments have successively introduced a series of policies and measures to promote the development of scientific and technological innovation, and technology innovation policies has played a very important leading role.Quantitative analysis of the technology innovation policies can enable policy makers to understand the current status of policies and adjust them. At the same time, it can also enable scientific research institutions, enterprises, individuals and other groups related to technology innovation policies to understand the focus and development trend of policies, so as to better carry out the relevant industry activities of technology innovation and integrate them into the national development strategy. However, there are many conceptual elements involved in technology innovation policies, the semantic relationship between conceptual elements is complex, and the traditional policy measurement method is relatively simple, which is difficult to comprehensively and deeply measure the characteristics of each conceptual element and the relationship between elements in technology innovation policy. Therefore, a more efficient way of organizing knowledge on technology innovation policies is needed, so that more correlations between conceptual elements of technology innovation policies and the information contained in them can be better explored through policy measurement.Based on the above demand analysis, in order to better understand the technology innovation policies and make a more comprehensive and efficient policy quantitative analysis, this paper introduces the knowledge graph technology, dedicates itself to building a technology innovation policy graph containing the main internal and external characteristics and connotation of technology innovation policy, so as to realize the semantic organization of technology innovation policy text resources, and explores the application mode of technology innovation policy graph in the field of policy measurement. In order to achieve this goal, this paper divides the technology innovation policy graph into concept layer, logic layer and data layer, and builds it successively. First of all, after absorbing theories related to technology innovation such as Triple Helix Theory, Policy Tool Theory and Innovation Value Chain Theory, and combining with China's actual situation, this paper analyzes the elements structure of technology innovation policy, and builds an analysis model of the technology innovation policies including the external characteristics of the technology innovation policies, policy tools, the subjects of the activities of  technology innovation and the stage of innovation value chain and other core elements. This model is taken as the conceptual layer of the technology innovation policy graph to provide theoretical and conceptual basis for the construction of the technology innovation policy graph. Then, on the basis of this policy analysis model, the ontology conceptual model of technology innovation policy is constructed as the knowledge representation basis of the technology innovation policy graph. The technology innovation policies of Shenzhen and Hong Kong, two core cities in the Guangdong-Hong Kong-Macao Greater Bay Area, are taken as data samples. The knowledge of technology innovation policy is extracted according to the concept structure of the ontology model, and the ontology instance is created, then the logical layer of technology innovation policy graph is successfully constructed. Next, according to the mapping relationship between the elements of the ontology model and the elements of the graph database, the extracted knowledge is stored in Neo4j graph database to complete the construction of data layer and realizes the construction of technology innovation policy graph. Then, two main application models of technology innovation policy graph in the field of policy measurement are proposed, namely, the exploration of the relations of policy measurement based on semantic relation reasoning and the retrieval of quantitative information based on semantic relation. And the quantitative relation between the elements of technology innovation policy that can be used for policy measurement and the method and process of retrieving the information needed for measurement by the technology innovation policy graph are sorted out. Finally, this paper takes Shenzhen and Hong Kong as the analyze objects, and selects five quantitative relations based on semantic relation reasoning as the measurement perspective, and completes multi-dimensional policy quantitative analysis based on information retrieval relaid on semantic relation, verifying the practicability and feasibility of the application models of the technology innovation policy graph in the field of policy measurement.