学术资源推荐服务中多维度用户画像构建  

Construction of Multi-dimensional User Profile in Academic Resource Recommendation Service

作　　者：顾亦然 郭玉雯 GU Yi-ran;GUO Yu-wen(School of Automation and Artificial Intelligence,Nanjing University of Posts and Telecommunications,Nanjing 210046,China)

机构地区：[1]南京邮电大学自动化学院、人工智能学院,江苏南京210046

出　　处：《软件导刊》2021年第4期101-105,共5页Software Guide

基　　金：国防基础科研项目(JCKY2019210B005,JCKY2018204B025,JCKY2017204B011);国防重大工程项目(ZQ2019D20401);军委装备发展部仿真预研项目(41401030301);江苏省研究生实践创新计划项目(SJCX19_0262)。

摘　　要：针对学术资源推荐系统优化,提出一种包含3个维度的学术用户画像,在用户基本信息和行为特征的基础上,融入“学术人格”这一概念。首先通过调查问卷的方式获取真实的用户数据,然后构建多维度学术用户画像模型,最后以主观评价的方式对基于用户画像的推荐结果进行分析验证。结果显示,多维度用户画像完善了学术用户画像的理论体系,优化了学术资源推荐服务策略。The academic personas combined with three dimensions is proposed for the optimization of academic resource recommenda⁃tion system.Based on the attributes and behavior of users,the concept of academic personality is integrated into the personas.Firstly,a questionnaire survey is made to obtain the user data and construct the multi-dimensional user profile.Then the recommendation re⁃sults based on user profile are analyzed and verified by subjective evaluation.The results shows that the multi-dimensional user profile improves the theoretical system of academic user profile,and optimizies the strategy of academic resources recommendation service.

关 键 词：用户画像 推荐系统 多维度 学术用户 学术人格 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]