面向舆情知识图谱的演化图数据事件查询与分析技术

Public sentiment knowledge graph is one type of graph data which built on public sentiment information. The query and analysis technique on public sentiment knowledge is an important technical method to query and analyze public opinion sentiment. According to user-defined search contents and analysis rules, it returns few valuable public sentiment events to users. Compared with the traditional graph data query and analysis techniques, it focuses on the existing relationships and the potential correlations between the events. Due to the huge scale of public sentiment knowledge graph and has the characteristics of multi-source heterogeneity, dynamic evolution and significant timeliness, which also increases the difficulty of data management and query analysis on public sentiment graph. In this project, we address to study some new techniques for supporting event query and analysis on public sentiment knowledge graph. In order to achieve this goal, we will study three important problems, which are public sentiment knowledge graph model construction, efficient data management and query analysis. In the first problem, we will study the techniques of heterogeneous data transformation and event relationship measurement. For the second one, we will study on the efficiently distributed storage and index structures of public sentiment knowledge graph. For the third problem, we aim to achieve the goal of querying and analyzing targets quickly. Finally, we plan to implement an oriented prototype system. The research results based on this project will provide a more reliable guarantee for improving the public sentiment graph, which has very important theoretical significance and practical application value.

舆情知识图谱是根据舆情信息构建的一类图谱数据。舆情知识图谱上的查询与分析是检索和分析舆情信息的重要技术手段，其根据用户自定义的检索内容和分析规则，返回有价值的舆情事件信息。和传统的图数据查询与分析技术相比，更注重事件之间已有的和潜在的关联关系。由于舆情知识图谱规模巨大，且具有多源异构性、动态演化性和显著时效性等特点，对舆情知识图谱的数据管理和查询分析带来巨大挑战。课题从舆情知识图谱建模、数据管理和查询分析三方面，对面向舆情知识图谱的演化图数据事件查询与分析技术展开深入研究。研究舆情知识图谱数据模型构建，实现异构数据的转化和事件关系度量；研究舆情知识图谱的分布式存储与索引结构，实现舆情数据的高效存储与索引；研究舆情知识图谱的查询与分析技术，实现事件查询与分析结果的快速返回。课题最终将实现一个原型系统，相关研究成果将对舆情事件查询与分析服务提供技术保证，具有十分重要的理论意义和实际应用价值。

舆情知识图谱上的事件查询与分析研究具有重要的应用价值。舆情知识图谱和传统的图数据查询与分析技术相比，更注重事件之间已有的和潜在的关联关系。由于舆情知识图谱规模巨大，且具有多源异构性、动态演化性和显著时效性等特点，对舆情知识图谱的数据管理和查询分析带来巨大挑战。项目从舆情知识图谱建模、数据管理和查询分析三方面，对面向舆情知识图谱的演化图数据事件查询与分析技术展开深入研究。研究舆情知识图谱数据模型构建，实现异构数据的转化和事件关系度量；研究舆情知识图谱的分布式存储与索引结构，实现舆情数据的高效存储与索引；研究舆情知识图谱的查询与分析技术，实现事件查询与分析结果的快速返回。项目组开展一系列相关的研究工作，提出了多项支持舆情知识图谱事件查询与分析相关技术和方法，撰写并发表论文11篇，其中期刊论文8篇，会议论文3篇，部分研究成果正在进行实际应用推广中。