| 社会化标注系统标签质量影响因素研究:基于随机森林算法 |
| |
| 引用本文： | 张云中,秦艺源.社会化标注系统标签质量影响因素研究:基于随机森林算法[J].图书情报工作,2019,63(24):119-126. |
| |
| 作者姓名： | 张云中 秦艺源 |
| |
| 作者单位： | 上海大学图书情报档案系 上海 200444 |
| |
| 基金项目： | 本文系国家哲学社会科学基金项目"基于形式概念分析的社会化标注系统语义发现与语义映射研究"（项目编号：16CTQ023）研究成果之一。 |
| |
| 摘 要： | 目的/意义] 在社会化标注系统中,标签质量往往关乎用户对网络资源的分类、查询、浏览、获取等使用体验,确定影响标签质量的关键因素有助于进一步优化社会化标注系统的资源组织核心功能。方法/过程] 以社会化标注系统的标签为研究对象,从标注主体、标注客体、标注环境、标注动机、标注方式、标注产物等维度入手重构标签质量影响因素模型,尝试探究影响社会化标签质量的关键因素,并运用问卷调查方法收集数据,结合有监督学习的随机森林算法,建立标签质量影响因素的决策树模型。结果/结论] 结果显示,标注主体是影响标签质量的首要关键维度,主体的知识结构和认知水平、标注频度及其感知有用性对标签质量的影响突出;标注方式是影响标签质量的次要关键维度,标签推荐和规范标签提示是影响标签质量的重要因素。
|
| 关 键 词： | 社会化标注系统 标签质量 机器学习 随机森林 |
| 收稿时间： | 2019-04-28 |
Research on Influencing Factors of Tag Quality in Social Tagging System: Based on Random Forest |
| |
| Zhang Yunzhong,Qin Yiyuan.Research on Influencing Factors of Tag Quality in Social Tagging System: Based on Random Forest[J].Library and Information Service,2019,63(24):119-126. |
| |
| Authors: | Zhang Yunzhong Qin Yiyuan |
| |
| Institution: | Department of Library, Information and Archives, Shanghai University, Shanghai 200444 |
| |
| Abstract: | Purpose/significance] Tag quality is often related to users' experience of classification, query, browsing, acquisition of online resources in social tagging system. Identifying key influencing factors of tag quality can optimize the core functions of resources organization of STS.Method/process] Based on tags, we provided the influencing factors model of tag quality from six perspectives, which covered tagging subject, tagging object, tagging environment, tagging motivation, tagging methods and tagging products. The study attempted to explore the key influencing factors of tag quality by questionnaire, and established the decision tree model of influencing factors of tag quality based on Random Forest.Result/conclusion] Tagging subject is the primary key dimension affecting tag quality. And the impact of the subject's knowledge structure and cognitive level, the subject's tagging frequency, and the subject's perceived usefulness are prominent. Tagging methods are the secondary one, and tag recommendation and standard tag tips are main influencing factors. |
| |
| Keywords: | social tagging system tag quality machine learning random forest |
|
| 点击此处可从《图书情报工作》浏览原始摘要信息 |
| 点击此处可从《图书情报工作》下载免费的PDF全文 |
|
