基于学术博客综合指数的用户识别方法及其实证研究  

User Identification Methods and Empirical Study Based on Composite Index of Academic Blogs

作　　者：袁润[1] 王琦 Yuan Run;Wang Qi(Institute of Science and Technology Information,Jiangsu University,Zhenjiang 212013,China)

机构地区：[1]江苏大学科技信息研究所

出　　处：《图书情报研究》2020年第1期93-100,共8页Library and Information Studies

基　　金：国家社会科学基金项目“图书馆知识发现服务的功能定位和建设策略研究”(项目编号:14BTQ018)的研究成果之一

摘　　要：[目的/意义]为深入开展学术博客用户研究,提出“博客综合指数”(BCSI)的概念,构建用户识别模型。[方法/过程]利用Python语言编写程序获取科学网博客用户行为数据并构建指标体系,采用熵权法确定指标权重系数并计算博客综合指数。[结果/结论]基于用户行为数据构建的指标体系能够表征用户主要特征,BCSI具有较好的区分度,可应用于用户识别。实证研究表明,科学网博客的BCSI值总体偏小,反映用户行为偏好差异性较大,规范性不好。分组研究表明,不同学科用户行为具有不同的特征。所构建的指标体系均属正向性指标,所以依据BCSI值的大小可以间接地评价学术博客影响力。[Purpose/significance]In order to carry out researches on academic blog users,this paper proposes the concept of blog composite index and constructs an user identification model.[Method/process]This research uses Python to capture the behavior data of ScienceNet users and construct an user identification index system.Entropy weight method is used to determine the index weight coefficient and calculate the blog composite index.[Result/conclusion]It is found that the index system based on users'behavior data can represent the main characteristics of users.BCSI has a good discrimination and can be applied to user identification.The empirical research shows that the BCSI value of the ScienceNet is on the low side,reflecting the difference in user behavior preferences and the poor standardization.Group study shows that user behavior of different disciplines shows different characteristics.The index system is constructed by positive indicator,so the impact of academic blog can be indirectly evaluated according to BCSI value.

关 键 词：学术博客 博客综合指数 用户识别 R语言 

分 类 号：G350[文化科学—情报学]