基于机器学习的学术论文推荐方法研究  

Researches on Recommendation Method of Academic Thesis Based on Machine Learning

作　　者：魏斌 万福成[1,2] 于洪志[1,2] 马宁[1,2] 杨恒 WEI Bin;WAN Fu-cheng;YU Hong-zhi;MA Ning;YANG Heng(Key Laboratory of China's Ethnic Languages and Information Technology of Ministry of Education,Lanzhou 730000,China;Key Laboratory of Intelligent Processing of Ethnic Languages in Gansu Province,Lanzhou 730000,China;Documentation and Information Centre,Northwest Institute of Eco-Environment and Resources,Chinese Academy of Sciences,Lanzhou 730030,China)

机构地区：[1]西北民族大学中国民族语言文字信息技术教育部重点实验室,甘肃兰州730030 [2]西北民族大学甘肃省民族语言智能处理重点实验室,甘肃兰州730030 [3]中国科学院西北生态环境资源研究院文献情报中心,甘肃兰州730030

出　　处：《西北民族大学学报（自然科学版）》2023年第3期72-83,共12页Journal of Northwest Minzu University(Natural Science)

基　　金：西北民族大学中央高校资助项目(31920230004)。

摘　　要：目前,学术论文的数量呈指数增长,论文推荐也已成为一项有吸引力的研究.论文推荐系统具有一定的重要性和优势.文章通过调查已有的一些论文推荐方法,如基于协同过滤的、图的、混合等方法,并对已有方法进行分析和总结的基础上,指出了目前学术论文推荐研究面临的挑战,以期探索出解决挑战的新思路、新方法.Due to the exponential growth in the number of academic theses,the thesis recommendation has become an attractive area of research.This article first introduced the importance and advantages of paper recommendation system.Secondly,it reviewed some existing methods used in thesis recommendation,such as content-based,collaborative filtering,graph-based and hybrid methods.Finally,it introduced the commonly used evaluation methods and academic thesis data sets,and pointed out the challenges facing the current researches on the academic thesis recommendation based on the analysis and summary of the existing methods.

关 键 词：学术论文推荐 推荐算法 协同过滤 图结构 

分 类 号：TP317.1[自动化与计算机技术—计算机软件与理论]