基于用户行为数据的学术用户画像构建方法研究
Research on the Construction Method of Academic User Profile Based on User Behavior Data
张良 1肖银涛1
作者信息
- 1. 同方知网数字出版技术股份有限公司,北京 100192
- 折叠
摘要
阐述了一种基于用户行为数据的学术用户画像构建方法,包括标记用户行为数据并按照三个维度标记用户身份;收集、清洗用户行为数据、用户访问过的学术资源的特征信息;构建用户兴趣领域和每个兴趣领域的关键词向量表达;结合用户账号相关信息构建学术用户画像.能够基于用户IP、账号、终端标识三个维度的历史行为数据,通过挖掘分析相关学术资源特征信息,构建学术用户画像.其中基于终端的学术用户画像不依赖于用户账号体系,为后续的机构读者个性化知识推荐服务提供支撑.
Abstract
This paper elaborates a method for constructing academic user profiles based on user behavior data,including labeling user behavior data and labeling user identity according to three dimensions.It collects and clean user behavior data and characteristic information of academic resources accessed by users,constructs vector representations of user interest domains and key words of each interest domain,and constructs an academic user profile based on user account related information.Based on historical behavioral data from three dimensions of user IP,account,and terminal identification,academic user profiles can be constructed by mining and analyzing relevant academic resource characteristic information.The academic user profile based on terminals does not rely on the user account system,providing support for personalized knowledge recommendation services for institutional readers in the future.
关键词
用户画像/个性化推荐/知识服务/学术资源/用户行为Key words
user profile/personalized recommendation/knowledge service/academic resource/user behavior引用本文复制引用
出版年
2024
国家科技期刊平台
NSTL
万方数据