基于相互增强和知识继承的多实体动态异构学术网络构建及应用  

CONSTRUCTION AND APPLICATION OF MULTI-ENTITY DYNAMIC HETEROGENEOUS ACADEMIC NETWORK BASED ON MUTUAL ENHANCEMENT AND KNOWLEDGE INHERITANCE

作　　者：马建红 宋秦汉 Ma Jianhong;Song Qinhan(School of Artificial Intelligence,Hebei University of Technology,Tianjin 300401,China)

机构地区：[1]河北工业大学人工智能与数据科学学院,天津300401

出　　处：《计算机应用与软件》2024年第6期287-295,共9页Computer Applications and Software

摘　　要：对科学家、研究组织和研究资助机构来说,准确客观地衡量作者的学术成就、评估论文和地点的学术水平是一项至关重要且具有挑战性的任务。各种基于图的排名方法,如PageRank已经被广泛用于在同构网络中对作者、论文和地点进行排名,但是仅限于在同构网络中解决这个问题,不适用于异构网络。为此,基于作者、论文和地点三种类型的实体构建一个多实体动态异构学术网络,并提出一种新的模型MEKI-Rank。模型中提出知识继承的概念,即作者通过其他作者继承知识,同时提取出动态异构网络中的七种关系,基于这些关系进行迭代排名,并使用每一轮的结果来相互增强排名。Accurate and objective measurement of the academic achievements of authors and assessment of the academic quality of papers and sites is a vital and challenging task for scientists,research organizations and research funding agencies.Various graph-based ranking methods,such as PageRank,have been widely used to rank authors,papers and places in homogeneous networks,but they are limited to solve this problem in homogeneous networks and are not applicable to heterogeneous networks.Therefore,this paper constructed a multi-entity dynamic heterogeneous academic network based on three entities of author,paper and place,and proposed a new model,MEKI-Rank.In this model,the concept of knowledge inheritance was proposed,that was,the author inherited knowledge from other authors.It extracted seven relationships in the dynamic heterogeneous network at the same time,carried out iterative ranking based on these relationships,and used the results of each round to enhance the ranking of each other.

关 键 词：多实体动态异构学术网络 知识继承 相互增强 迭代排名 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]