CAJ | 학술논문

目前，聚合服务包含网络资源聚合、服务聚合与用户聚合三个方面，其中用户聚合研究亟待完善。在一些开放平台的社会化标签系统中，用户根据个人偏好自由选择词汇对网络资源进行标注。标注的标签词反映了用户的兴趣偏好。从标签共现网络视角，提出了一种基于标签共现网络的用户聚合算法TBKM。定义了TBKM算法的相关概念；创新提出了TBKM算法并给出了算法的详细设计；选择目前网络上最大的书签类站点Delicious的真实数据进行实证研究，将TBKM算法与传统K-means算法在聚类效果上进行比较，展示了TBKM算法在簇间距离与簇内半径两指标上的创新性。
목전，취합복무포함망락자원취합、복무취합여용호취합삼개방면，기중용호취합연구극대완선。재일사개방평태적사회화표첨계통중，용호근거개인편호자유선택사회대망락자원진행표주。표주적표첨사반영료용호적흥취편호。종표첨공현망락시각，제출료일충기우표첨공현망락적용호취합산법TBKM。정의료TBKM산법적상관개념；창신제출료TBKM산법병급출료산법적상세설계；선택목전망락상최대적서첨류참점Delicious적진실수거진행실증연구，장TBKM산법여전통K-means산법재취류효과상진행비교，전시료TBKM산법재족간거리여족내반경량지표상적창신성。
Aggregation services include resource aggregation, service aggregation and user aggregation, whereas only few researches focus on user aggregation. In some open platform systems of social tagging, user can tag resource with freely chose vocabulary in terms of their own need. The words in tags reflect user’s preference. This paper proposes a user clustering algorithm named TBKM from the perspective of social tags co-occurrence net. The paper is organized as follows. It defines some concepts with regard to TBKM. It elaborates on the novel algorithm TBKM. An experiment on real data from a book tag website named Delicious shows the difference in effectiveness between TBKM and original K-means.