基于知识图谱的电网科技评审专家推荐方法研究  

Research on Expert Recommendation Method for Power Grid Science and Technology Review Based on Knowledge Graph

作　　者：徐翀 魏冠元 王冠群 刘紫熹 王其清 张炎 XU Chong;WEI Guanyuan;WANG Guanqun;LIU Zixi;WANG Qiqing;ZHANG Yan

机构地区：[1]国网能源研究院有限公司,北京102209 [2]国网智联电商有限公司,天津300000

出　　处：《青海电力》2022年第3期67-72,共6页Qinghai Electric Power

基　　金：国家电网公司科技项目“基于知识图谱的科技咨询专家智能优选技术研究与开发”(1400-202057269A-0-0-00)。

摘　　要：为了提高电网科技项目咨询评审专家遴选效率,提出基于知识图谱的电网科技专家推荐方法。首先构建包含专家、项目和研究领域的电网科技知识图谱,再通过TF-IDF算法提取专家论文和项目领域的关键词,基于音形码编码规则对关键词进行重新编译并通过编辑距离映射至图谱节点,最后利用余弦相似度计算得到专家与项目的最佳匹配组合。以某电网科技项目评审专家推荐为算例,通过K近邻法验证了所提方法的有效性,可在一定程度上增强电网科技项目评审专家遴选质量与效率。In order to improve the selection efficiency of power grid science and technology project consulting and review experts,this paper proposes a power grid science and technology expert recommendation method based on knowledge graph.Firstly,construct a power grid science and technology knowledge graph including experts,projects and research fields,and then extract the keywords of expert papers and project fields through the TF-IDF algorithm,recompile the keywords based on the phonetic code coding rules,and map them to the graph through edit distance.Finally,the cosine similarity calculation is used to obtain the best matching combination of experts and projects.Taking the recommendation of a power grid science and technology project review expert as an example,the K-nearest neighbor method is used to verify the effectiveness of the proposed method,which can enhance the selection quality and efficiency of power grid science and technology project review experts to a certain extent.

关 键 词：知识图谱 推荐算法 音形码 相似度计算 

分 类 号：TP391.1[自动化与计算机技术—计算机应用技术]